-*- mode: org -*-
Medicine, biology, biomechanics, bioinformatics, and chemistry (incl. process technology)
*  15(<-604): Consensus feature selection for multi-objective SVM modeling of protein ion-exchange displacement chromatography.



2005


*  38(<-327): MHC-I prediction using a combination of T cell epitopes and MHC-I binding peptides

We propose a novel learning method that combines multiple experimental
modalities to improve the MHC Class-I binding prediction. Multiple
experimental modalities are often accessible in the context of a
binding problem. Such modalities can provide different labels of data,
such as binary classifications, affinity measurements, or direct
estimations of the binding profile. Current machine learning
algorithms usually focus on a given label type. We here present a
novel Multi-Label Vector Optimization (MEMO) formalism to produce
classifiers based on the simultaneous optimization of multiple
labels. Within this methodology, all label types are combined into a
single constrained quadratic dual optimization problem. We apply the
MLVO to MHC class-I epitope prediction. We combine affinity
measurements (IC50/EC50), binary classifications of epitopes as T cell
activators and existing algorithms. The multi-label vector
optimization algorithms produce classifiers significantly better than
the ones resulting from any of its components. These matrix based
classifiers are better or equivalent to the existing state of the art
MHC-I epitope prediction tools in the studied alleles. (C) 2010
Elsevier B.V. All rights reserved.

2011



*  55(<-174): Multiobjective Binary Biogeography Based Optimization for Feature Selection Using Gene Expression Data

Gene expression data play an important role in the development of
efficient cancer diagnoses and classification. However, gene
expression data are usually redundant and noisy, and only a subset of
them present distinct profiles for different classes of samples. Thus,
selecting high discriminative genes from gene expression data has
become increasingly interesting in the field of bioinformatics. In
this paper, a multi-objective biogeography based optimization method
is proposed to select the small subset of informative gene relevant to
the classification. In the proposed algorithm, firstly, the
Fisher-Markov selector is used to choose the 60 top gene expression
data. Secondly, to make biogeography based optimization suitable for
the discrete problem, binary biogeography based optimization, as
called BBBO, is proposed based on a binary migration model and a
binary mutation model. Then, multi-objective binary biogeography based
optimization, as we called MOBBBO, is proposed by integrating the
non-dominated sorting method and the crowding distance method into the
BBBO framework. Finally, the MOBBBO method is used for gene selection,
and support vector machine is used as the classifier with the
leave-one-out cross-validation method (LOOCV). In order to show the
effective and efficiency of the algorithm, the proposed algorithm is
tested on ten gene expression dataset benchmarks. Experimental results
demonstrate that the proposed method is better or at least comparable
with previous particle swarm optimization (PSO) algorithm and support
vector machine (SVM) from literature when considering the quality of
the solutions obtained.

2013



*  56(<-254): Multiclass Gene Selection Using Pareto-Fronts

Filter methods are often used for selection of genes in multiclass sample classification by using microarray data. Such techniques usually tend to bias toward a few classes that are easily distinguishable from other classes due to imbalances of strong features and sample sizes of different classes. It could therefore lead to selection of redundant genes while missing the relevant genes, leading to poor classification of tissue samples. In this manuscript, we propose to decompose multiclass ranking statistics into class-specific statistics and then use Pareto-front analysis for selection of genes. This alleviates the bias induced by class intrinsic characteristics of dominating classes. The use of Pareto-front analysis is demonstrated on two filter criteria commonly used for gene selection: F-score and KW-score. A significant improvement in classification performance and reduction in redundancy among top-ranked genes were achieved in experiments with both synthetic and real-benchmark data sets.

2013





*  58(<-321): MicroRNA transcription start site prediction with multi-objective feature selection.

MicroRNAs (miRNAs) are non-coding, short (21-23nt) regulators of protein-coding genes that are generally transcribed first into primary miRNA (pri-miR), followed by the generation of precursor miRNA (pre-miR). This finally leads to the production of the mature miRNA. A large amount of information is available on the pre- and mature miRNAs. However, very little is known about the pri-miRs, due to a lack of knowledge about their transcription start sites (TSSs). Based on the genomic loci, miRNAs can be categorized into two types --intragenic (intra-miR) and intergenic (inter-miR). While it is already an established fact that intra-miRs are commonly transcribed in conjunction with their host genes, the transcription machinery of inter-miRs is poorly understood. Although it is assumed that miRNA promoters are similar in structure to gene promoters, since both are transcribed by RNA polymerase II (Pol II), computational validations exhibit poor performance of gene promoter prediction methods on miRNAs. In this paper, we concentrate on the problem of TSS prediction for miRNAs. The present study begins with the identification of positive and negative promoter samples from recently published data stemming from RNA-sequencing studies. From these samples of experimentally validated miRNA TSSs, a number of standard sequence features are extracted. Furthermore, to account for potential footprints related to promoter regulation by CpG dinucleotide targeted DNA methylation, a number of novel features are defined. We develop a support vector machine (SVM) with RBF kernel for the prediction of miRNA TSSs trained on human miRNA promoters. A novel feature reduction technique based on archived multi-objective simulated annealing (AMOSA) identifies the final set of features. The resulting model trained on miRNA promoters shows improved performance over the one trained on protein-coding gene promoters in terms of classification accuracy, sensitivity and specificity. Results are also reported for a completely independent biologically validated test set. In a part of the investigation, the proposed approach is used to predict protein-coding gene TSSs. It shows a significantly improved performance when compared to previously published gene TSS prediction methods.

2012



*  59(<-337): MultiMiTar: A Novel Multi Objective Optimization based miRNA-Target Prediction Method

Background: Machine learning based miRNA-target prediction algorithms often fail to obtain a balanced prediction accuracy in terms of both sensitivity and specificity due to lack of the gold standard of negative examples, miRNA-targeting site context specific relevant features and efficient feature selection process. Moreover, all the sequence, structure and machine learning based algorithms are unable to distribute the true positive predictions preferentially at the top of the ranked list; hence the algorithms become unreliable to the biologists. In addition, these algorithms fail to obtain considerable combination of precision and recall for the target transcripts that are translationally repressed at protein level. Methodology/Principal Finding: In the proposed article, we introduce an efficient miRNA-target prediction system MultiMiTar, a Support Vector Machine (SVM) based classifier integrated with a multiobjective metaheuristic based feature selection technique. The robust performance of the proposed method is mainly the result of using high quality negative examples and selection of biologically relevant miRNA-targeting site context specific features. The features are selected by using a novel feature selection technique AMOSA-SVM, that integrates the multi objective optimization technique Archived Multi-Objective Simulated Annealing (AMOSA) and SVM. Conclusions/Significance: MultiMiTar is found to achieve much higher Matthew's correlation coefficient (MCC) of 0.583 and average class-wise accuracy (ACA) of 0.8 compared to the others target prediction methods for a completely independent test data set. The obtained MCC and ACA values of these algorithms range from -0.269 to 0.155 and 0.321 to 0.582, respectively. Moreover, it shows a more balanced result in terms of precision and sensitivity (recall) for the translationally repressed data set as compared to all the other existing methods. An important aspect is that the true positive predictions are distributed preferentially at the top of the ranked list that makes MultiMiTar reliable for the biologists. MultiMiTar is now available as an online tool at www.isical.ac.in/similar to bioinfo similar to miu/multimitarhtm. MultiMiTar software can be downloaded from www.isical.ac.in/similar to bioinfo_miu/multimitar-download.htm..

2011



*  60(<-547): Gene selection with multiple ordering criteria

Background: A microarray study may select different differentially expressed gene sets because of different selection criteria. For example, the fold-change and p-value are two commonly known criteria to select differentially expressed genes under two experimental conditions. These two selection criteria often result in incompatible selected gene sets. Also, in a two-factor, say, treatment by time experiment, the investigator may be interested in one gene list that responds to both treatment and time effects. Results: We propose three layer ranking algorithms, point-admissible, line-admissible (convex), and Pareto, to provide a preference gene list from multiple gene lists generated by different ranking criteria. Using the public colon data as an example, the layer ranking algorithms are applied to the three univariate ranking criteria, fold-change, p-value, and frequency of selections by the SVM-RFE classifier. A simulation experiment shows that for experiments with small or moderate sample sizes (less than 20 per group) and detecting a 4-fold change or less, the two-dimensional (p-value and fold-change) convex layer ranking selects differentially expressed genes with generally lower FDR and higher power than the standard p-value ranking. Three applications are presented. The first application illustrates a use of the layer rankings to potentially improve predictive accuracy. The second application illustrates an application to a two-factor experiment involving two dose levels and two time points. The layer rankings are applied to selecting differentially expressed genes relating to the dose and time effects. In the third application, the layer rankings are applied to a benchmark data set consisting of three dilution concentrations to provide a ranking system from a long list of differentially expressed genes generated from the three dilution concentrations. Conclusion: The layer ranking algorithms are useful to help investigators in selecting the most promising genes from multiple gene lists generated by different filter, normalization, or analysis methods for various objectives.

2007



*  61(<-173): An SVM-Wrapped Multiobjective Evolutionary Feature Selection Approach for Identifying Cancer-MicroRNA Markers

MicroRNAs (miRNAs), have been shown to play important roles in gene regulation and various biological processes. Recent studies have revealed that abnormal expression of some specific miRNAs often results in the development of cancer. Microarray datasets containing the expression profiles of several miRNAs are being used for identification of miRNAs which are differentially expressed in normal and malignant tissue samples. In this article, a multiobjective feature selection approach is proposed for this purpose. The proposed method uses Genetic Algorithm for multiobjective optimization and support vector machine (SVM) classifier as a wrapper for evaluating the chromosomes that encode feature subsets. The performance has been demonstrated on real-life miRNA datasets for and the identified miRNA markers are reported. Moreover biological significance tests have been carried out for the obtained markers.

2013



*  62(<-381): Gene expression data analysis using multiobjective clustering improved with SVM based ensemble.

Microarray technology facilitates the monitoring of the expression levels of thousands of genes over different experimental conditions simultaneously. Clustering is a popular data mining tool which can be applied to microarray gene expression data to identify co-expressed genes. Most of the traditional clustering methods optimize a single clustering goodness criterion and thus may not be capable of performing well on all kinds of datasets. Motivated by this, in this article, a multiobjective clustering technique that optimizes cluster compactness and separation simultaneously, has been improved through a novel support vector machine classification based cluster ensemble method. The superiority of MOCSVMEN (MultiObjective Clustering with Support Vector Machine based ENsemble) has been established by comparing its performance with that of several well known existing microarray data clustering algorithms. Two real-life benchmark gene expression datasets have been used for testing the comparative performances of different algorithms. A recently developed metric, called Biological Homogeneity Index (BHI), which computes the clustering goodness with respect to functional annotation, has been used for the comparison purpose.

2011



*  63(<-392): Multi-Class Clustering of Cancer Subtypes through SVM Based Ensemble of Pareto-Optimal Solutions for Gene Marker Identification

With the advancement of microarray technology, it is now possible to study the expression profiles of thousands of genes across different experimental conditions or tissue samples simultaneously. Microarray cancer datasets, organized as samples versus genes fashion, are being used for classification of tissue samples into benign and malignant or their subtypes. They are also useful for identifying potential gene markers for each cancer subtype, which helps in successful diagnosis of particular cancer types. In this article, we have presented an unsupervised cancer classification technique based on multiobjective genetic clustering of the tissue samples. In this regard, a real-coded encoding of the cluster centers is used and cluster compactness and separation are simultaneously optimized. The resultant set of near-Pareto-optimal solutions contains a number of non-dominated solutions. A novel approach to combine the clustering information possessed by the non-dominated solutions through Support Vector Machine (SVM) classifier has been proposed. Final clustering is obtained by consensus among the clusterings yielded by different kernel functions. The performance of the proposed multiobjective clustering method has been compared with that of several other microarray clustering algorithms for three publicly available benchmark cancer datasets. Moreover, statistical significance tests have been conducted to establish the statistical superiority of the proposed clustering method. Furthermore, relevant gene markers have been identified using the clustering result produced by the proposed clustering method and demonstrated visually. Biological relationships among the gene markers are also studied based on gene ontology. The results obtained are found to be promising and can possibly have important impact in the area of unsupervised cancer classification as well as gene marker identification for multiple cancer subtypes.

2010



*  64(<-484): Combining Pareto-optimal clusters using supervised learning for identifying co-expressed genes

Background: The landscape of biological and biomedical research is being changed rapidly with the invention of microarrays which enables simultaneous view on the transcription levels of a huge number of genes across different experimental conditions or time points. Using microarray data sets, clustering algorithms have been actively utilized in order to identify groups of co-expressed genes. This article poses the problem of fuzzy clustering in microarray data as a multiobjective optimization problem which simultaneously optimizes two internal fuzzy cluster validity indices to yield a set of Pareto-optimal clustering solutions. Each of these clustering solutions possesses some amount of information regarding the clustering structure of the input data. Motivated by this fact, a novel fuzzy majority voting approach is proposed to combine the clustering information from all the solutions in the resultant Pareto-optimal set. This approach first identifies the genes which are assigned to some particular cluster with high membership degree by most of the Pareto-optimal solutions. Using this set of genes as the training set, the remaining genes are classified by a supervised learning algorithm. In this work, we have used a Support Vector Machine (SVM) classifier for this purpose. Results: The performance of the proposed clustering technique has been demonstrated on five publicly available benchmark microarray data sets, viz., Yeast Sporulation, Yeast Cell Cycle, Arabidopsis Thaliana, Human Fibroblasts Serum and Rat Central Nervous System. Comparative studies of the use of different SVM kernels and several widely used microarray clustering techniques are reported. Moreover, statistical significance tests have been carried out to establish the statistical superiority of the proposed clustering approach. Finally, biological significance tests have been carried out using a web based gene annotation tool to show that the proposed method is able to produce biologically relevant clusters of co-expressed genes. Conclusion: The proposed clustering method has been shown to perform better than other well-known clustering algorithms in finding clusters of co-expressed genes efficiently. The clusters of genes produced by the proposed technique are also found to be biologically significant, i.e., consist of genes which belong to the same functional groups. This indicates that the proposed clustering method can be used efficiently to identify co-expressed genes in microarray gene expression data. Supplementary Website The pre-processed and normalized data sets, the matlab code and other related materials are available at http://anirbanmukhopadhyay.50webs.com/mogasvm.html.

2009





*  67(<-148): Multi-objective evolutionary algorithms for fuzzy classification in survival prediction

Objective: This paper presents a novel rule-based fuzzy classification
methodology for survival/mortality prediction in severe burnt
patients. Due to the ethical aspects involved in this medical
scenario, physicians tend not to accept a computer-based evaluation
unless they understand why and how such a recommendation is
given. Therefore, any fuzzy classifier model must be both accurate and
interpretable. Methods and materials: The proposed methodology is a
three-step process: (1) multi-objective constrained optimization of a
patient's data set, using Pareto-based elitist multi-objective
evolutionary algorithms to maximize accuracy and minimize the
complexity (number of rules) of classifiers, subject to
interpretability constraints; this step produces a set of alternative
(Pareto) classifiers; (2) linguistic labeling, which assigns a
linguistic label to each fuzzy set of the classifiers; this step is
essential to the interpretability of the classifiers; (3) decision
making, whereby a classifier is chosen, if it is satisfactory,
according to the preferences of the decision maker. If no classifier
is satisfactory for the decision maker, the process starts again in
step (1) with a different input parameter set. Results: The
performance of three multi-objective evolutionary algorithms, niched
pre-selection multi-objective algorithm, elitist Pareto-based
multi-objective evolutionary algorithm for diversity reinforcement
(ENORA) and the non-dominated sorting genetic algorithm (NSGA-II), was
tested using a patient's data set from an intensive care burn unit and
a standard machine learning data set from an standard machine learning
repository. The results are compared using the hypervolume
multi-objective metric. Besides, the results have been compared with
other non-evolutionary techniques and validated with a multi-objective
cross-validation technique. Our proposal improves the classification
rate obtained by other non-evolutionary techniques (decision trees,
artificial neural networks, Naive Bayes, and case-based reasoning)
obtaining with ENORA a classification rate of 0.9298, specificity of
0.9385, and sensitivity of 0.9364, with 14.2 interpretable fuzzy rules
on average. Conclusions: Our proposal improves the accuracy and
interpretability of the classifiers, compared with other
non-evolutionary techniques. We also conclude that ENORA outperforms
niched pre-selection and NSGA-II algorithms. Moreover, given that our
multi-objective evolutionary methodology is non-combinational based on
real parameter optimization, the time cost is significantly reduced
compared with other evolutionary approaches existing in literature
based on combinational optimization. (C) 2014 Elsevier B.V. All rights
reserved.

2014





*  74(<-602): Ensemble strategies for a medical diagnostic decision support system: A breast cancer diagnosis application

The model selection strategy is an important determinant of the
performance and acceptance of a medical diagnostic decision support
system based on supervised learning algorithms. This research
investigates the potential of various selection strategies from a
population of 24 classification models to form ensembles in order to
increase the accuracy of decision support systems for the early
detection and diagnosis of breast cancer. Our results suggest that
ensembles formed from a diverse collection of models are generally
more accurate than either pure-bagging ensembles (formed from a single
model) or the selection of a "single best model." We find that
effective ensembles are formed from a small and selective subset of
the population of available models with potential candidates
identified by a multicriteria process that considers the properties of
model generalization error, model instability, and the independence of
model decisions relative to other ensemble members. (C) 2003 Elsevier
B.V. All rights reserved.

2005





*  76(<-635): A novel hybrid GA/SVM system for protein sequences classification

A novel hybrid genetic algorithm(GA)/Support Vector Machine (SVM) system, which selects features from the protein sequences and trains the SVM classifier simultaneously using a multi-objective genetic algorithm, is proposed in this paper. The system is then applied to classify protein sequences obtained from the Protein Information Resource (PIR) protein database. Finally, experimental results over six protein superfamilies are reported, where it is shown that the proposed hybrid GA/SVM system outperforms BLAST and HMMer.

2004





*  79(<- 13): Multi-kernel multi-criteria optimization classifier with fuzzification and penalty factors for predicting biological activity

Nowadays it is important to develop effective computational methods
for accurately identifying and predicting biological activity in the
virtual screening of bioassay data so as to speed up the process of
drug development. Among these methods, multi-criteria optimization
classifier (MCOC) is a classifier which can find a trade-off between
the overlapping degree of different classes and the total distance
from input points to the decision hyperplane. The former should be
minimized while the latter should be maximized. Then a decision
function is derived from training data and this function is
subsequently used to predict the class label of an unseen
instance. However, due to outliers, anomalies, highly imbalanced
classes, high dimension, nonlinear separability and other
uncertainties in data, MCOC and other methods often give the poor
predictive performance. In this paper, we introduce a new fuzzy
contribution to each input point based on class median, by defining
the new row and column kernel functions the linear combination of
different feature kernels to replace the single kernel function in the
kernel-induced feature space and penalty factors to imbalanced
classes, thus a novel multi-kernel multi-criteria optimization
classifier with fuzzification and penalty factors (MK-MCOC-FP) is
proposed and the effects of the aforementioned problems are
significantly reduced. The experimental results of predicting active
compounds in the virtual screening and comparison with linear and
quadratic MCOCs, support vector machines (SVM), fuzzy SVM and neural
network, the conclusions show that MK-MCOC-FP evidently increased the
ability of resisting noise interference, the predictive accuracy of
highly class-imbalanced bioassay data, the separation of active
compounds and inactive compounds, the interpretability of importance
or contributions of different features to classification, the
efficiency of classification with feature selection or dimensionality
reduction for high-dimensional data, and the generalization of
predicting the biological activity of new compounds. (C) 2015 Elsevier
B.V. All rights reserved.

2015




*  81(<-132): Multi-criteria optimization classifier using fuzzification, kernel and penalty factors for predicting protein interaction hot spots

In order to understand the patterns of various biological processes
and discover the principles of protein-protein interactions (PPI), it
is important to develop effective methods for identifying and
predicting PPI and their hot spots accurately. As for multi-criteria
optimization classifier (MCOC), it can learn a decision function from
different classes of training data and use it to predict the class
labels of unknown samples. In many real world applications, owing to
noises, outliers, imbalanced class distribution, nonlinearly separable
problems, and other uncertainties, the predictive performance of MCOC
degenerates rapidly. In this paper, we introduce a fuzzy contribution
to each instance of training data, the unequal penalty factors to the
samples in imbalanced classes, and kernel method to nonlinearly
separable dataset, then a novel multi-criteria optimization classifier
with fuzzification, kernel and penalty factors (FKP-MCOC) is
constructed so as to reduce the effects of anomalies, improve the
class imbalanced performance, and nonlinear separability in
classification. The experimental results of predicting active
compounds and protein interaction hot spots and comparison with MCOC,
support vector machines (SVM) and fuzzy SVM, the conclusion shows that
FKP-MCOC significantly increases the efficiency of classification, the
partition of active and inactive compounds in bioassay, the separation
of hot spot residues and energetically unimportant residues in protein
interactions, and the generalization of predicting active compounds
and hot spot residues in new instances. (C) 2014 Elsevier B.V. All
rights reserved.

2014





* 105(<-408): De novo design: balancing novelty and confined chemical space

Importance of the field: De novo drug design serves as a tool for the
discovery of new ligands for macromolecular targets as well as
optimization of known ligands. Recently developed tools aim to address
the multi-objective nature of drug design in an unprecedented
manner. Areas covered in this review: This article discusses recent
advances in de novo drug design programs and accessory programs used
to evaluate compounds post-generation. What the reader will gain: The
reader is introduced to the challenges inherent in de novo drug design
and will become familiar with current trends in de novo
design. Furthermore, the reader will be better prepared to assess the
value of a tool, and be equipped to design more elegant tools in the
future. Take home message: De novo drug design can assist in the
efficient discovery of new compounds with a high affinity for a given
target. The inclusion of existing chemoinformatic methods with current
structure-based de novo design tools provides a means of enhancing the
therapeutic value of these generated compounds.

2010




* 143(<-506): Combined effect of solvent content, temperature and pH on the chromatographic behaviour of ionisable compounds - II: Benefits of the simultaneous optimisation

A previously reported eight-parameter mechanistic model [Part I of
this work,J. Chromatogr. A 1163 (2007) 49] was applied to optimise the
separation of 11 ionisable compounds (nine diuretics and two
beta-blockers), considering solvent content, temperature and pH as
experimental factors. The data from 21 experiments, arranged in a
central composite design, were used to model the retention. Local
models were used to predict efficiency and peak asymmetry. The
optimisation strategy, based on the use of peak purity as
chromatographic objective function and derived concepts, was able to
find the most suitable experimental conditions yielding full
resolution in reasonable analysis times. It also allowed a detailed
inspection of the separation capability of the studied factors, and of
the consequences of the shifts in the protonation constants originated
by changes in solvent content and temperature. The size of the
resolution structures suggested that the ranked importance of the
factors was pH, organic solvent and temperature, giving rise to
relatively narrow domains of full resolution. The three factors were
found, however, worthwhile in the optimisation of
selectivity. Predicted optimal conditions corresponding to two
different optimal resolution regions were verified experimentally. In
spite of the difficulties associated to the use of pH as optimisation
factor, satisfactory agreement was found in both cases. (c) 2008
Elsevier B.V. All rights reserved.

2008





* 183(<-618): Feasibility study of beam orientation class-solutions for prostate IMRT

IMRT is being increasingly used for treatment of prostate cancer. In
practice, however, the beam orientations used for the treatments are
still selected empirically, without any guideline. The purpose of this
work was to investigate interpatient variation of the optimal beam
configuration and to facilitate intensity modulated radiation therapy
(IMRT) prostate treatment planning by proposing a set of beam
orientation class-solutions for a range of numbers of incident
beams. We used fifteen prostate cases to generate the beam orientation
class-solutions. For each patient and a given number of incident
beams, a multiobjective optimization engine was employed to provide
optimal beam directions. For the fifteen cases considered, the gantry
angle of any of the optimized plans were all distributed within a
certain range The angular distributions of the optimal beams were
analyzed and the most selected directions are identified as optimal
directions. The optimal directions for all patients are averaged to
obtain the class-solution. The class-solution gantry angles for
prostate IMRT were found to be: three beams (0degrees, 120degrees,
240degrees), five beams (35degrees, 110degrees, 180degrees,
250degrees, 325degrees), six beams (0degrees, 60degrees, 120degrees,
180degrees, 240degrees, 300degrees), seven beams (25degrees,
75degrees, 130degrees, 180', 230degrees, 285degrees, 335degrees),
eight beams (20degrees, 70degrees, 110degrees, 150degrees, 200degrees,
250degrees, 290degrees, 340degrees), and nine beams (20degrees,
60degrees, 100degrees, 140degrees, 180degrees, 220degrees, 260degrees,
300degrees, 340degrees). The level of validity of the class-solutions
was tested using an additional clinical prostate case by comparing
with the individually optimized beam configurations. The difference
between the plans obtained with class-solutions and patient-specific
optimizations was found to be clinically insignificant. (C) 2004
American Association of Physicists in Medicine.

2004




* 221(<-442): On the possibility of non-invasive multilayer temperature estimation using soft-computing methods

Objective and motivation: This work reports original results on the
possibility of non-invasive temperature estimation (NITE) in a
multilayered phantom by applying soft-computing methods. The existence
of reliable non-invasive temperature estimator models would improve
the security and efficacy of thermal therapies. These points would
lead to a broader acceptance of this kind of therapies. Several
approaches based on medical imaging technologies were proposed,
magnetic resonance imaging (MRI) being appointed as the only one to
achieve the acceptable temperature resolutions for hyperthermia
purposes. However, MRI intrinsic characteristics (e. g., high
instrumentation cost) lead us to use backscattered ultrasound
(BSU). Among the different BSU features, temporal echo-shifts have
received a major attention. These shifts are due to changes of
speed-of-sound and expansion of the medium. Novelty aspects: The
originality of this work involves two aspects: the estimator model
itself is original (based on soft-computing methods) and the
application to temperature estimation in a three-layer phantom is also
not reported in literature. Materials and methods: In this work a
three-layer (non-homogeneous) phantom was developed. The two external
layers were composed of (in % of weight): 86.5% degassed water, 11%
glycerin and 2.5% agar agar. The intermediate layer was obtained by
adding graphite powder in the amount of 2% of the water weight to the
above composition. The phantom was developed to have attenuation and
speed-of-sound similar to in vivo muscle, according to the
literature. BSU signals were collected and cumulative temporal
echo-shifts computed. These shifts and the past temperature values
were then considered as possible estimators inputs. A soft-computing
methodology was applied to look for appropriate multilayered
temperature estimators. The methodology involves radial-basis
functions neural networks (RBFNN) with structure optimized by the
multi-objective genetic algorithm (MOGA). In this work 40 operating
conditions were considered, i.e. five 5-mm spaced spatial points and
eight therapeutic intensities (I(SATA)): 0.3, 0.5, 0.7, 1.0, 1.3, 1.5,
1.7 and 2:0 W/cm(2). Models were trained and selected to estimate
temperature at only four intensities, then during the validation
phase, the best-fitted models were analyzed in data collected at the
eight intensities. This procedure leads to a more realistic evaluation
of the generalisation level of the best-obtained structures. Results
and discussion: At the end of the identification phase, 82
(preferable) estimator models were achieved. The majority of them
present an average maximum absolute error (MAE) inferior to 0.5
degrees C. The best-fitted estimator presents a MAE of only 0.4
degrees C for both the 40 operating conditions. This means that the
gold-standard maximum error (0.5 degrees C) pointed for hyperthermia
was fulfilled independently of the intensity and spatial position
considered, showing the improved generalisation capacity of the
identified estimator models. As the majority of the preferable
estimator models, the best one presents 6 inputs and 11 neurons. In
addition to the appropriate error performance, the estimator models
present also a reduced computational complexity and then the
possibility to be applied in real-time. Conclusions: A non-invasive
temperature estimation model, based on soft-computing technique, was
proposed for a three-layered phantom. The best-achieved estimator
models presented an appropriate error performance regardless of the
spatial point considered (inside or at the interface of the layers)
and of the intensity applied. Other methodologies published so far,
estimate temperature only in homogeneous media. The main drawback of
the proposed methodology is the necessity of a-priory knowledge of the
temperature behavior. Data used for training and optimisation should
be representative, i.e., they should cover all possible physical
situations of the estimation environment. (C) 2009 Elsevier B.V. All
rights reserved.

2010


* 222(<-577): Non-invasive temperature prediction of in vitro therapeutic ultrasound signals using neural networks

In this paper, a novel black-box modelling scheme applied to non-invasive temperature prediction in a homogeneous medium subjected to therapeutic ultrasound is presented. It is assumed that the temperature in a point of the medium is non-linearly related to some spectral features and one temporal feature, extracted from the collected RF-lines. The black-box models used are radial basis functions neural networks (RBFNNs), where the best-fitted models were selected from the space of model structures using a genetic multi-objective strategy. The best-fitted predictive model presents a maximum absolute error less than 0.4 degrees C in a prediction horizon of approximately 2 h, in an unseen data sequence. This work demonstrates that this type of black-box model is well-suited for punctual and non-invasive temperature estimation, achieving, for a single point estimation, better results than the ones presented in the literature, encouraging research on multi-point non-invasive temperature estimation.

2006




* 236(<-371): Two-dimensional fingerprinting approach for comparison of complex substances analysed by HPLC-UV and fluorescence detection

This work is concerned with the research and development of
methodology for analysis of complex mixtures such as pharmaceutical or
food samples, which contain many analytes. Variously treated samples
(swill washed, fried and scorched) of the Rhizoma atractylodis
macrocephalae (RAM) traditional Chinese medicine (TCM) as well as the
common substitute, Rhizoma atractylodis (RA) TCM were chosen as
examples for analysis. A combined data matrix of chromatographic 2-D
HPLC-DAD-FLD (two-dimensional high performance liquid chromatography
with diode array and fluorescence detectors) fingerprint profiles was
constructed with the use of the HPLC-DAD and HPLC-FLD individual data
matrices; the purpose was to collect maximum information and to
interpret this complex data with the use of various chemometrics
methods e. g. the rank-ordering multi-criteria decision making (MCDM)
PROMETHEE and GAIA, K-nearest neighbours (KNN), partial least squares
(PLS), back propagation-artificial neural networks (BP-ANN)
methods. The chemometrics analysis demonstrated that the combined 2-D
HPLC-DAD-FLD data matrix does indeed provide more information and
facilitates better performing classification/prediction models for the
analysis of such complex samples as the RAM and RA ones noted
above. It is suggested that this fingerprint approach is suitable for
analysis of other complex, multi-analyte substances.

2011





* 264(<-221): Predicting patient survival after liver transplantation using evolutionary multi-objective artificial neural networks

Objective: The optimal allocation of organs in liver transplantation is a problem that can be resolved using machine-learning techniques. Classical methods of allocation included the assignment of an organ to the first patient on the waiting list without taking into account the characteristics of the donor and/or recipient. In this study, characteristics of the donor, recipient and transplant organ were used to determine graft survival. We utilised a dataset of liver transplants collected by eleven Spanish hospitals that provides data on the survival of patients three months after their operations. Methods and material: To address the problem of organ allocation, the memetic Pareto evolutionary non-dominated sorting genetic algorithm 2 (MPENSGA2 algorithm), a multi-objective evolutionary algorithm, was used to train radial basis function neural networks, where accuracy was the measure used to evaluate model performance, along with the minimum sensitivity measurement. The neural network models obtained from the Pareto fronts were used to develop a rule-based system. This system will help medical experts allocate organs. Results: The models obtained with the MPENSGA2 algorithm generally yielded competitive results for all performance metrics considered in this work, namely the correct classification rate (C), minimum sensitivity (MS), area under the receiver operating characteristic curve (AUC), root mean squared error (RMSE) and Cohen's kappa (Kappa). In general, the multi-objective evolutionary algorithm demonstrated a better performance than the mono-objective algorithm, especially with regard to the MS extreme of the Pareto front, which yielded the best values of MS (48.98) and AUC (0.5659). The rule-based system efficiently complements the current allocation system (model for end-stage liver disease, MELD) based on the principles of efficiency and equity. This complementary effect occurred in 55% of the cases used in the simulation. The proposed rule-based system minimises the prediction probability error produced by two sets of models (one of them formed by models guided by one of the objectives (entropy) and the other composed of models guided by the other objective (MS)), such that it maximises the probability of success in liver transplants, with success based on graft survival three months post-transplant. Conclusion: The proposed rule-based system is objective, because it does not involve medical experts (the expert's decision may be biased by several factors, such as his/her state of mind or familiarity with the patient). This system is a useful tool that aids medical experts in the allocation of organs; however, the final allocation decision must be made by an expert. (C) 2013 Elsevier B.V. All rights reserved.

2013




* 267(<-263): Multi-objective evolutionary algorithm for donor-recipient decision system in liver transplants

This paper reports on a decision support system for assigning a liver from a donor to a recipient on a waiting-list that maximises the probability of belonging to the survival graft class after a year of transplant and/or minimises the probability of belonging to the non-survival graft class in a two objective framework. This is done with two models of neural networks for classification obtained from the Pareto front built by a multi-objective evolutionary algorithm - called MPENSGA2. This type of neural network is a new model of the generalised radial basis functions for obtaining optimal values in C (Correctly Classified Rate) and MS (Minimum Sensitivity) in the classifier, and is compared to other competitive classifiers. The decision support system has been proposed using, as simply as possible, those models which lead to making the correct decision about receptor choice based on efficient and impartial criteria. (c) 2012 Elsevier B.V. All rights reserved.

2012



* 272(<-583): Differentiation of syndromes with SVM

Differentiation of syndromes is the kernel theory of Traditional
Chinese Medicine (TCM). How to diagnose syndromes correctly with
scientific means according to symptoms is the first problem in
TCM. Several modem approaches have been applied, but no satisfied
results have been obtained because of the complexity of diagnosis
procedure. Support Vector Machine (SVM) is a new classification
technique and has drawn much attention on this topic in recent
years. In this paper, we combine non-linear Principle Component
Analysis (PCA) neural network with multi-class SVM to realize
differentiation of syndromes. Non-linear PCA is used to preprocess
clinical data to save computational cost and reduce noise. The
multi-class SVM takes the non-linear principle components as its
inputs and determines a corresponding syndrome. Analyzing of a TCM
example shows its effectiveness.

2006




* 273(<- 72): Multi-Criteria Decision Making: The Best Choice for the Modeling of Chemicals Against Hyper-Pigmentation?

Classifier ensembles appeared to be powerful alternative for handling
a difficult problem. It is rapidly growing and enjoying many
attentions from pattern recognition and machine learning
communities. In the present report, the potential of multi-criteria
decision making via multiclassifier approaches is assessed by applying
them in the modeling of chemicals against
hyper-pigmentation. TOMOCOMD-CARDD atom-based quadratic indices are
used as descriptors to parameterize the molecular structures. Support
vector machine, artificial neural network, Bayesian network, binary
logistic regression, instance-based learning and tree classification
applied on two collected datasets are explored as standalone
classifiers. Prediction sets (PSs) are used to assess the performance
of multiclassifier systems (MCSs). A strategy exploiting the principal
component analysis together with pairwise diversity measures is
designed to select the most diverse base classifiers to
combine. Various trainable and nontrainable systems are developed that
aggregate, at the abstract and continuous levels, the outputs of base
classifiers. The obtained results are rather encouraging since the
MCSs generally enhance the performance of the base classifiers;
e.g. the best MCS obtains global accuracy of 95.51%, 88.89% in the PS
for the data I and II in regard to 94.12% and 85.93% of best
individual classifier, respectively. Our results suggest that the MCSs
could be the best choice till the moment to obtain suitable QSAR
models for the prediction of depigmenting agents. Finally, we consider
these approaches will aid improving the virtual screening procedures
and increasing the practicality of data mining of chemical datasets
for the discovery of novel lead compounds.

2015




* 281(<-464): A hybrid multi-objective genetic algorithm for evaluation of essential sets of medical diagnostic factors

A hybrid algorithm that incorporates two biologically inspired
computational intelligence methods was used for the assessment of
abdominal pain. Namely, Genetic Algorithms (GA) where used in search
for the optimal subset of clinical diagnostic factors that can be
given as inputs for Probabilistic Neural Networks (PNN) to perform
medical diagnosis based on the clinical data. Thus, the implemented GA
was a two-objective one. The first objective was to minimize the
number of diagnostic factors that were considered for medical
diagnosis. The second objective was to minimize the Mean Square Error
of the constructed PNN at the testing phase. The obtained results of
the proposed hybrid algorithm are related favorably to the
corresponding ones derived by applying Receiver Operating
Characteristic analysis. Eventually, it was found that a number up to
60% of the diagnostic factors that are recorded in patient's history
may be omitted without any loss in clinical assessment validity,
while, at the same time, the performance of the genetically pruned PNN
is improved in terms of execution speed and prediction accuracy.

2009




* 289(<-248): A Clinical Decision Support System for Femoral Peripheral Arterial Disease Treatment

One of the major challenges of providing reliable healthcare services is to diagnose and treat diseases in an accurate and timely manner. Recently, many researchers have successfully used artificial neural networks as a diagnostic assessment tool. In this study, the validation of such an assessment tool has been developed for treatment of the femoral peripheral arterial disease using a radial basis function neural network (RBFNN). A data set for training the RBFNN has been prepared by analyzing records of patients who had been treated by the thoracic and cardiovascular surgery clinic of a university hospital. The data set includes 186 patient records having 16 characteristic features associated with a binary treatment decision, namely, being a medical or a surgical one. K-means clustering algorithm has been used to determine the parameters of radial basis functions and the number of hidden nodes of the RBFNN is determined experimentally. For performance evaluation, the proposed RBFNN was compared to three different multilayer perceptron models having Pareto optimal hidden layer combinations using various performance indicators. Results of comparison indicate that the RBFNN can be used as an effective assessment tool for femoral peripheral arterial disease treatment.

2013




* 290(<-457): Optimal construction of a fast and accurate polarisable water potential based on multipole moments trained by machine learning

To model liquid water correctly and to reproduce its structural,
dynamic and thermodynamic properties warrants models that account
accurately for electronic polarisation. We have previously
demonstrated that polarisation can be represented by fluctuating
multipole moments (derived by quantum chemical topology) predicted by
multilayer perceptrons (MLPs) in response to the local structure of
the cluster. Here we further develop this methodology of modeling
polarisation enabling control of the balance between accuracy, in
terms of errors in Coulomb energy and computing time. First, the
predictive ability and speed of two additional machine learning
methods, radial basis function neural networks (RBFNN) and Kriging,
are assessed with respect to our previous MLP based polarisable water
models, for water dimer, trimer, tetramer, pentamer and hexamer
clusters. Compared to MLPs, we find that RBFNNs achieve a 14-26%
decrease in median Coulomb energy error, with a factor 2.5-3 slowdown
in speed, whilst Kriging achieves a 40-67% decrease in median energy
error with a 6.5-8.5 factor slowdown in speed. Then, these compromises
between accuracy and speed are improved upon through a simple
multi-objective optimisation to identify Pareto-optimal
combinations. Compared to the Kriging results, combinations are found
that are no less accurate (at the 90th energy error percentile), yet
are 58% faster for the dimer, and 26% faster for the pentamer.

2009





* 291(<-440): Application of Different Artificial Neural Networks Retention Models for Multi-Criteria Decision-Making Optimization in Gradient Ion Chromatography

In this work, the principles of multi-criteria decision-making were
used to develop an efficient optimization strategy in gradient elution
ion chromatographic analysis. Two different artificial neural network
retention models (multi-layer perceptron and radial basis function),
three different separation criterion functions (chromatography
response function, separation factor product and normalized retention
difference product), and four different robustness criterion functions
(CR1-CR4) were examined. The shape of the calculated separation vs the
robustness response surface was used as principal criterion. Analysis
time and minimum separation of adjacent peaks were additional
criteria. The results showed that the radial basis artificial neural
network retention model in combination with normalized retention
difference product separation criterion function and CR3 robustness
criterion function provided the optimal gradient ion chromatographic
analysis.

2010




* 296(<- 58): A new training algorithm using artificial neural networks to classify gender-specific dynamic gait patterns

The aim of this study was to present a new training algorithm using
artificial neural networks called multi-objective least absolute
shrinkage and selection operator (MOBJ-LASSO) applied to the
classification of dynamic gait patterns. The movement pattern is
identified by 20 characteristics from the three components of the
ground reaction force which are used as input information for the
neural networks in gender-specific gait classification. The
classification performance between MOBJ-LASSO (97.4%) and
multi-objective algorithm (MOBJ) (97.1%) is similar, but the
MOBJ-LASSO algorithm achieved more improved results than the MOBJ
because it is able to eliminate the inputs and automatically select
the parameters of the neural network. Thus, it is an effective tool
for data mining using neural networks. From 20 inputs used for
training, MOBJ-LASSO selected the first and second peaks of the
vertical force and the force peak in the antero-posterior direction as
the variables that classify the gait patterns of the different
genders.

2015





* 298(<-189): A genetic algorithm-based multi-objective optimization of an artificial neural network classifier for breast cancer diagnosis

The conventional technique for diagnosing the breast cancer disease
relies on human experiences to identify the presence of certain
pattern from the database. It is time-consuming and incurs unnecessary
burden to radiologists. This work proposes a genetic algorithm-based
multi-objective optimization of an Artificial Neural Network
classifier, namely GA-MOO-NN, for the automatic breast cancer
diagnosis. It performs a simultaneous search for the significant
feature subsets and the optimum architecture of the network. The
combination of ANN's parameters with feature selection to be optimized
by Genetic Algorithm is novel. The Pareto-optimality with new ranking
approach is applied for simultaneous minimizations of two competing
objectives: the number of network's connections and squared error
percentage of the validation data. Result shows that the algorithm
with the proposed combination of objectives has achieved the best and
average, 98.85 and 98.10 % accuracy of classification, respectively,
on breast cancer dataset which outperform most systems of other works
found in the literature.

2013




* 299(<-280): AMS 4.0: consensus prediction of post-translational modifications in protein sequences

We present here the 2011 update of the AutoMotif Service (AMS 4.0) that predicts the wide selection of 88 different types of the single amino acid post-translational modifications (PTM) in protein sequences. The selection of experimentally confirmed modifications is acquired from the latest UniProt and Phospho.ELM databases for training. The sequence vicinity of each modified residue is represented using amino acids physico-chemical features encoded using high quality indices (HQI) obtaining by automatic clustering of known indices extracted from AAindex database. For each type of the numerical representation, the method builds the ensemble of Multi-Layer Perceptron (MLP) pattern classifiers, each optimising different objectives during the training (for example the recall, precision or area under the ROC curve (AUC)). The consensus is built using brainstorming technology, which combines multi-objective instances of machine learning algorithm, and the data fusion of different training objects representations, in order to boost the overall prediction accuracy of conserved short sequence motifs. The performance of AMS 4.0 is compared with the accuracy of previous versions, which were constructed using single machine learning methods (artificial neural networks, support vector machine). Our software improves the average AUC score of the earlier version by close to 7 % as calculated on the test datasets of all 88 PTM types. Moreover, for the selected most-difficult sequence motifs types it is able to improve the prediction performance by almost 32 %, when compared with previously used single machine learning methods. Summarising, the brainstorming consensus meta-learning methodology on the average boosts the AUC score up to around 89 %, averaged over all 88 PTM types. Detailed results for single machine learning methods and the consensus methodology are also provided, together with the comparison to previously published methods and state-of-the-art software tools. The source code and precompiled binaries of brainstorming tool are available at http://code.google.com/p/automotifserver/under Apache 2.0 licensing.

2012




* 305(<-351): Donor-Recipient Matching Based on a Rule-System Built on a Multiobjective Artificial Neural Network



2011


* 306(<-361): Donor-Recipient Matching in Liver Transplantation Based on a Rule-System Built on a Multiobjective Artificial Neural Network



2011






* 319(<-243): Memetic Pareto differential evolutionary neural network used to solve an unbalanced liver transplantation problem

Donor-recipient matching constitutes a complex scenario difficult to
model. The risk of subjectivity and the likelihood of falling into
error must not be underestimated. Computational tools for the
decision-making process in liver transplantation can be useful,
despite the inherent complexity involved. Therefore, a multi-objective
evolutionary algorithm and various techniques to select individuals
from the Pareto front are used in this paper to obtain artificial
neural network models to aid decision making. Moreover, a combination
of two pre-processing methods has been applied to the dataset to
offset the existing imbalance. One of them is a resampling method and
the other is a outlier deletion method. The best model obtained with
these procedures (with AUC = 0.66) give medical experts a probability
of graft survival at 3 months after the operation. This probability
can help medical experts to achieve the best possible decision without
forgetting the principles of fairness, efficiency and equity.

2013





* 345(<-620): Classification of HIV-1-mediated neuronal dendritic and synaptic damage using multiple criteria linear programming

The ability to identify neuronal damage in the dendritic arbor during
HIV-1-associated dementia (HAD) is crucial for designing specific
therapies for the treatment of HAD. To study this process, we utilized
a computer-based image analysis method to quantitatively assess HIV-1
viral protein gp120 and glutamate-mediated individual neuronal damage
in cultured cortical neurons. Changes in the number of neurites,
arbors, branch nodes, cell body area, and average arbor lengths were
determined and a database was formed
(http://dm.st.unomaha. edu/database.htm). We further proposed a
two-class model of multiple criteria linear programming (MCLP) to
classify such HIV-1-mediated neuronal dendritic and synaptic
damages. Given certain classes, including treatments with
brain-derived neurotrophic factor (BDNF), glutamate, gp120 or
non-treatment controls from our in vitro experimental systems, we used
the two-class MCLP model to determine the data patterns between
classes in order to gain insight about neuronal dendritic
damages. This knowledge can be applied in principle to the design and
study of specific therapies for the prevention or reversal of neuronal
damage associated with HAD. Finally, the MCLP method was compared with
a well-known artificial neural network algorithm to test for the
relative potential of different data mining applications in HAD
research.

2004





* 350(<-507): Neuro-genetic non-invasive temperature estimation: Intensity and spatial prediction

Objectives: The existence of proper non-invasive temperature
estimators is an essential aspect when thermal therapy applications
are envisaged. These estimators must be good predictors to enable
temperature estimation at different operational situations, providing
better control of the therapeutic instrumentation. In this work,
radial basis functions artificial neural networks were constructed to
access temperature evolution on an ultrasound insonated
medium. Methods: The employed models were radial basis functions
neural networks with external dynamics induced by their inputs. Both
the most suited set of model inputs and number of neurons in the
network were found using the multi-objective genetic algorithm. The
neural models were validated in two situations: the operating ones, as
used in the construction of the network; and in 11 unseen
situations. The new data addressed two new spatial locations and a new
intensity level, assessing the intensity and space prediction capacity
of the proposed model. Results: Good performance was obtained during
the validation process both in terms of the spatial points considered
and whenever the new intensity level was within the range of applied
intensities. A maximum absolute error of 0.5 degrees C +/- 10% (0.5
degrees C is the gold-standard threshold in hyperthermia/diathermia)
was attained with low computationally complex models. Conclusion: The
results confirm that the proposed neuro-genetic approach enables
foreseeing temperature propagation, in connection to intensity and
space parameters, thus enabling the assessment of different operating
situations with proper temperature resolution. (C) 2008 Elsevier
B.V. All rights reserved.

2008




* 353(<-194): Neural Networks Applied in Chemistry. II. Neuro-Evolutionary Techniques in Process Modeling and Optimization

Artificial neural networks are widely used in data analysis and to
control dynamic processes. These tools are powerful and versatile, but
the way in which they are constructed, in particular their
architecture, strongly affects their value and reliability. We review
here some key techniques for optimizing artificial neural networks and
comment on their use in process modeling and
optimization. Neuro-evolutionary techniques are described and
compared, with the goal of providing efficient modeling methodologies
which employ an optimal neural model. We also discuss how neural
networks and evolutionary algorithms can be combined. Applications
from chemical engineering illustrate the effectiveness and reliability
of the hybrid neuro-evolutionary methods.

2013




* 364(<-114): Multicriteria decision analysis: a multifaceted approach to medical equipment management

Selecting medical equipment is a complex multidisciplinary task
requiring mathematical tools, considering associated
uncertainties. This paper offers an in-depth study of
multiple-criteria decision analysis (MCDA) methods to identify the
most appropriate ones for performing management tasks in
resource-limited settings. The chosen articles were divided into three
topics: evaluation of projects and equipment, selection of projects
and equipment, and development of medical devices. Three methods
(analytic hierarchy process [AHP], multi-attribute utility theory and
elimination and choice expressing reality) were selected for detailed
analyses of their application for medical equipment
management. Twenty-one work using MCDA, artificial neural networks,
human factors engineering, and value analysis were analysed in the
framework of medical equipment management. The important aspects of
the procedure were described, highlighting their advantages and
disadvantages. It was determined that the AHP approach corresponds to
all defined criteria for selecting large medical equipment. Managing
large medical equipment using MCDA will reduce uncertainties, and
provide a rational selection and purchase of the most efficient
equipment in resource-limited settings. The direction for improving
the AHP method was determined.

2014



* 365(<-245): Gene expression rule discovery and multi-objective ROC analysis using a neural-genetic hybrid

Microarray data allows an unprecedented view of the biochemical
mechanisms contained within a cell although deriving useful
information from the data is still proving to be a difficult task. In
this paper, a novel method based on a multi-objective genetic
algorithm is investigated that evolves a near-optimal trade-off
between Artificial Neural Network (ANN) classifier accuracy
(sensitivity and specificity) and size (number of genes). This hybrid
method is shown to work on four well-established gene expression data
sets taken from the literature. The results provide evidence for the
rule discovery ability of the hybrid method and indicate that the
approach can return biologically intelligible as well as plausible
results and requires no pre-filtering or pre-selection of genes.

2013



* 366(<-600): A hybrid promoter analysis methodology for prokaryotic genomes

One of the big challenges of the post-genomic era is identifying regulatory systems and integrating them into genetic networks. Gene expression is determined by protein-protein interactions among regulatory proteins and with RNA polymerase(s), and protein-DNA interactions of these trans-acting factors with cis-acting DNA sequences in the promoter regions of those regulated genes. Therefore, identifying these protein-DNA interactions, by means of the DNA motifs that characterize the regulatory factors operating in the transcription of a gene, becomes crucial for determining, which genes participate in a regulation process, how they behave and how they are connected to build genetic networks. In this paper. we propose a hybrid promoter analysis methodology (HPAM) to discover complex promoter motifs that combines: the neural network efficiency and ability of representing imprecise and incomplete patterns; the flexibility and interpretability of fuzzy models; and the multi-objective evolutionary algorithms capability to identify optimal instances of a model by searching according to multiple criteria. We test our methodology by learning and predicting the RNA polymerase motif in prokaryotic genomes. This constitutes a special challenge due to the multiplicity of the RNA polymerase targets and its connectivity with other transcription factors, which sometimes require multiple functional binding sites even in close located regulatory regions; and the uncertainty of its motif, which allows sites with low specificity (i.e., differing from the best alignment or consensus) to still be functional. HPAM is available for public use in http://soar-tools.wustl.edu. (c) 2004 Elsevier B.V All rights reserved.

2005




* 373(<-454): [Artificial neural network parameters optimization software and its application in the design of sustained release tablets].

Artificial neural network (ANN) is a multi-objective optimization
method that needs mathematic and statistic knowledge which restricts
its application in the pharmaceutical research area. An artificial
neural network parameters optimization software (ANNPOS) programmed by
the Visual Basic language was developed to overcome this
shortcoming. In the design of a sustained release formulation, the
suitable parameters of ANN were estimated by the ANNPOS. And then the
Matlab 5.0 Neural Network Toolbox was used to determine the optimal
formulation. It showed that the ANNPOS reduced the complexity and
difficulty in the ANN's application.

2009



* 375(<-655): An evolutionary artificial neural networks approach for breast cancer diagnosis

This paper presents an evolutionary artificial neural network (EANN)
approach based on the pareto-differential evolution (PDE) algorithm
augmented with local search for the prediction of breast cancer. The
approach is named memetic pareto artificial neural network
(MPANN). Axtificial neural networks (ANNs) could be used to improve
the work of medical practitioners in the diagnosis of breast
cancer. Their abilities to approximate nonlinear functions and capture
complex relationships in the data are instrumental abilities which
could support the medical domain. We compare our resuits against an
evolutionary programming approach and standard backpropagation (BP),
and we show experimentally that MPANN has better generalization and
much lower computational cost. (C) 2002 Elsevier Science B.V. All
rights reserved.

2002




* 439(<-303): Multi-criteria decision making development of ion chromatographic method for determination of inorganic anions in oilfield waters based on artificial neural networks retention model

This paper describes the development of ad hoc methodology for
determination of inorganic anions in oilfield water, since their
composition often significantly differs from the average
(concentration of components and/or matrix). Therefore, fast and
reliable method development has to be performed in order to ensure the
monitoring of desired properties under new conditions. The method
development was based on computer assisted multi-criteria decision
making strategy. The used criteria were: maximal value of objective
functions used, maximal robustness of the separation method, minimal
analysis time, and maximal retention distance between two nearest
components. Artificial neural networks were used for modeling of anion
retention. The reliability of developed method was extensively tested
by the validation of performance characteristics. Based on validation
results, the developed method shows satisfactory performance
characteristics, proving the successful application of computer
assisted methodology in the described case study. (C) 2011 Elsevier
B.V. All rights reserved.

2012




* 495(<-403): A Neurogenetic Approach to a Multiobjective Design Optimization of Spinal Pedicle Screws

A pedicle screw fixation has been widely used to treat spinal
diseases. Clinical reports have shown that the weakest part of the
spinal fixator is the pedicle screw. However, previous studies have
only focused on either screw breakage or screw loosening. There have
been no studies that have addressed the multiobjective design
optimization of the pedicle screws. The multiobjective optimization
methodology was applied and it consisted of finite element method,
Taguchi method, artificial neural networks, and genetic
algorithms. Three-dimensional finite element models for both the
bending strength and the pullout strength of the pedicle screw were
first developed and arranged on an L(25) orthogonal array. Then,
artificial neural networks were used to create two objective
functions. Finally, the optimum solutions of the pedicle screws were
obtained by genetic algorithms. The results showed that the optimum
designs had higher bending and pullout strengths compared with
commercially available screws. The optimum designs of pedicle screw
revealed excellent biomechanical performances. The neuro genetic
approach has effectively decreased the time and effort required for
searching for the optimal designs of pedicle screws and has directly
provided the selection information to surgeons. [DOI:
10.1115/1.4001887]

2010



* 501(<-565): Evaluation of artificial neural networks for modelling and optimization of medium composition with a genetic algorithm

A new concept was evaluated for experimental multi-objective medium
optimization using a genetic algorithm which is supported by an
artificial neural network (ANN). The ANN is used to model objective
functions with the medium components as variables each time a new data
set has been produced. An appropriate topology of the ANN was first
identified with simulation studies using a multi-dimensional test
function (De Jong's function). The performance of this ANN model was
validated from generation to generation with the data of an
experimental optimization of a medium with 13 medium components for
Synechococcus PCC 7942. Objective functions were the simultaneous
maximization of biomass concentration and conversion of
pentafluoroacetophenon (PFAP) for asymmetric synthesis of
(S)-(-)1-(pentafluorophenyl)-ethanol. The mean absolute error of the
ANN simulation was within the experimental estimation error after six
from eight generations for one of the two objective functions (PFAP
conversion). This artificial neural network supported genetic
algorithm (ANNSGA) can thus be implemented at the end of a stochastic
optimization procedure to reduce the experimental effort. (c) 2006
Elsevier Ltd. All rights reserved.

2006




* 535(<-290): A Compact Optical Instrument with Artificial Neural Network for pH Determination

The aim of this work was the determination of pH with a sensor
array-based optical portable instrument. This sensor array consists of
eleven membranes with selective colour changes at different pH
intervals. The method for the pH calculation is based on the
implementation of artificial neural networks that use the responses of
the membranes to generate a final pH value. A multi-objective
algorithm was used to select the minimum number of sensing elements
required to achieve an accurate pH determination from the neural
network, and also to minimise the network size. This helps to minimise
instrument and array development costs and save on microprocessor
energy consumption. A set of artificial neural networks that fulfils
these requirements is proposed using different combinations of the
membranes in the sensor array, and is evaluated in terms of accuracy
and reliability. In the end, the network including the response of the
eleven membranes in the sensor was selected for validation in the
instrument prototype because of its high accuracy. The performance of
the instrument was evaluated by measuring the pH of a large set of
real samples, showing that high precision can be obtained in the full
range.

2012





* 552(<-273): A sensitive and efficient method for simultaneous trace detection and identification of triterpene acids and its application to pharmacokinetic study

A sensitive and efficient method for simultaneous trace detection of
seven triterpene acids was developed and validated for analysis of rat
plasma samples. The required micro-sampling of only 20 ut. blood
reduced the difficulty in blood collection and the injury to
animal. The whole pretreatment procedure was more conveniently
finished within 26 min through the application of the semi-automated
derivatization extraction method to biological samples. Seven analytes
were rapidly separated within 30 min on reversed-phase Akasil-C18
column and quantified by fluorescence detector. Online ion trap MS
with atmospheric pressure chemical ionization (APCI) source was used
for further identification. The novel application of artificial neural
network (ANN) combined with genetic algorithm (GA) to optimization of
derivatization was performed and compared with the classical response
surface methodology (RSM). Optimal derivatization condition was
validated by multi-criteria and nonparametric tests and used
successfully to achieve the rather high sensitivity (limit of
detection: 0.67-1.08 ng/mL). The limit of reactant concentration
(LORC) special for derivatization was studied and the lower values
(2.53-4.03 ng/mL) ensured the trace detection. Results of validation
demonstrated the advantages for pharmacokinetic study, such as higher
sensitivity, better accuracy, easier pretreatment and shorter
run-time. Pharmacokinetic study of triterpene acids after oral
administration of Salvia miltiorrhiza extract to mice was conducted
for the first time. The present method provided more sensitive and
efficient alternative for the medical detection of bioactive
constituents from herbal extract in the biological liquid. (C) 2012
Elsevier B.V. All rights reserved.

2012


* 553(<-276): A sensitive and efficient method to systematically detect two biophenols in medicinal herb, herbal products and rat plasma based on thorough study of derivatization and its convenient application to pharmacokinetics with semi-automated device

A sensitive and efficient method using a semi-automated pretreatment device, pre-column derivatization, multivariate optimization and high performance liquid chromatography with fluorescence and mass spectrometric detection was developed and validated for the systematic determination of two biophenols in four herb-related samples (medicinal herb; herbal products in tablet, capsule and oral liquid forms) and plasma samples after oral administration to rat. Only micro-sampling of 20 mu L blood was needed for the analysis, and the pretreatment procedure including blood collection, derivatization by 10-ethyl-acridine-3-sulfonyl chloride (EASC) and injection to the sampling vials was efficiently finished in 10 min with no cumbersome and complicated operation. The novel application of artificial neural network (ANN) coupled with genetic algorithm (GA) to optimization of derivatization condition was executed and compared with the classical response surface methodology (RSM). The optimal condition for derivatization was validated by multi-criteria and nonparametric tests and used successfully to achieve the higher sensitivity (limit of detection: 0.6 and 0.8 ng/mL). The limit of reactant concentration (LORC) was put forward for derivatization method for the first time, and the lower values (2.0-2.7 ng/mL) provided the guarantee for the trace detection with the micro samples (<50 mu L) required. The results of validation including selectivity, sensitivity, linearity, accuracy, precision, recovery, matrix effect and stability demonstrated the advantages of this method. The pharmacokinetic study of major bioactive components salidroside and p-tyrosol in herb Rhodiola crenulata and its products was more conveniently performed in 25 min. The established method could be the sensitive and efficient alternative method for the systematic detection of bioactive components in series of drug carriers from raw herb to herbal products and to blood in medical research. And the approaches of the thorough study played the guiding role in seeking a novel analytical method. (C) 2012 Elsevier B.V. All rights reserved.

2012





* 555(<-532): Modeling and optimization of m-cresol isopropylation for obtaining n-thymol: Combining a hybrid artificial neural network with a genetic algorithm

The application of a hybrid framework based on the combination,
artificial neural network-genetic algorithm (ANN-GA), for n-thymol
synthesis modeling and optimization has been developed. The effects of
molar ratio propylene/cresol (X1), catalyst mass (X2) and temperature
(X3) on n-thymol selectivity Y1 and m-cresol conversion Y2 were
studied. A 3-8-2 ANN model was found to be very suitable for reaction
modeling. The multiobjective optimization, led to optimal operating
conditions ( 0.55 <= X1 <= 0.77; 1.773 g <= X2 <= 1.86 g; 289.74
degrees C <= X3 <= 291.33 degrees C) representing good solutions for
obtaining high n-thymol selectivity and high m-cresol conversion. This
optimal zone corresponded to n-thymol selectivity and m-cresol
conversion ranging respectively in the interval [79.3; 79.5]% and [
13.4 %; 23.7]%. These results were better than those obtained with a
sequential method based on experimental design for which, optimum
conditions led to n-thymol selectivity and m-cresol conversion values
respectively equal to 67% and 11%. The hybrid method ANN-GA showed its
ability to solve complex problems with a good fitting.

2007




* 556(<- 97): Synthesis of ZnO nano-sono-catalyst for degradation of reactive dye focusing on energy consumption: operational parameters influence, modeling, and optimization

Simple synthesized Nano-sized ZnO powder in the absence of
high-temperature activation treatments was studied to act as
sono-catalyst. Effects of six operational parameters such as initial
solution pH (pH(0)), initial concentration of dye stuff (C-0),
additional dose of nano-sized ZnO powder (D-SC), ultrasound (US)
irradiation frequency (Fr-SC), US irradiation power (P-SC), and
treatment time (t(SC)) were examined. Synthetic wastewater containing
Reactive Red 198 (RR198) was used as the sample model. Combined design
of experiments was done and experiments were conducted according to
protocols. The experimental data were collected in a laboratory-scaled
batch reactor equipped with ultrasonic bath cleaner as the ultrasonic
source. The measured CR% ranging from 0.8 to 100 and EnC (wh) from 0.3
to 13.6 gained under given conditions. The data used for modeling were
used in two more common models in this type of studies: Multiple
linear regression (MLR) and artificial neural network (ANN). The ANN
models obviously outperformed MLR models. Finally, Multi-objective
optimization of CR% and EnC was carried out using genetic algorithm
(GA) over the outperformed ANN models. The optimization procedure
causes non-dominated optimal points which give an insight of the
optimal operating conditions.

2014



* 558(<-216): Optimization of OCM reactions over Na-W-Mn/SiO2 catalyst at elevated pressure using artificial neural network and response surface methodology

In this study, Response Surface Methodology (RSM) and Artificial
Neural Network (ANN) predictive models are developed, based on
experimental data of the Oxidative Coupling of Methane (OCM) over
Na-W-Mn/SiO2 at 0.4 MPa, which was obtained in an isothermal fixed bed
reactor. Results show that the simulation and prediction accuracy of
ANN was apparently higher compared to RSM. Thus, the Hybrid Genetic
Algorithm (HGA), based on developed ANN models, was used for
simultaneous maximization of CH4 conversion and C2+ selectivity. The
pareto optimal solutions show that at a reaction temperature of 987 K,
feed GHSV of 15790 h(-1), diluents amounts of 20 mole%, and methane to
oxygen molar ratio of 3.5, the maximum C2+ yield obtained from ANN-HGA
was 23.91% (CH4 conversion of 34.6% and C2+ selectivity of 69%), as
compared to 22.81% from the experimental measurements (CH4 conversion
of 34.0% and C2+ selectivity of 67.1%). The predicted error in optimum
yield by ANN-HGA was 4.81%, suggesting that the combination of ANN
models with the hybrid genetic algorithm could be used to find a
suitable operating condition for the OCM process at elevated
pressures. (C) 2013 Sharif University of Technology. Production and
hosting by Elsevier B.V. All rights reserved.

2013


* 559(<-339): Optimization of OCM reaction conditions over Na-W-Mn/SiO2 catalyst at elevated pressure

Performance of the oxidative coupling of methane (OCM) at elevated pressures has been simulated by a set of supervised Artificial Neural Network (ANN) models using reaction data gathered in a microreactor device. Accuracy of the developed models were evaluated by comparing the predicted results with the test data set showing a good agreement. In order to enhance the performance of OCM process at 0.4 MPa as a desired operating pressure for commercial application of OCM, the Hybrid Genetic Algorithm (HGA) was used to obtain the optimal values of the operating conditions. Nondominated Pareto optimal solutions were obtained and additional experiments were carried out at two different optimum conditions in order to verify the optimums. The results show that combination of ANN models with HGA could be used in finding the suitable operating conditions for OCM process at elevated pressures. It was shown that the C2+ yield of above 23% can be achieved at 0.4 MPa by using Na-W-Mn/SiO2 as the OCM catalyst. (C) 2011 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

2011




* 562(<-531): Modelling and optimization of catalytic-dielectric barrier discharge plasma reactor for methane and carbon dioxide conversion using hybrid artificial neural network - genetic algorithm technique

A hybrid artificial neural network-genetic algorithm (ANN-GA) was
developed to model, simulate and optimize the catalytic-dielectric
barrier discharge plasma reactor. Effects of CH4/CO2 feed ratio, total
feed flow rate, discharge voltage and reactor wall temperature on the
performance of the reactor was investigated by the ANN-based model
simulation. Pareto optimal solutions and the corresponding optimal
operating parameter range based on multi-objective scan be suggested
for two cases, i.e., simultaneous maximization of CH4 conversion and
C2+ selectivity (Case 1), and H-2 selectivity and H-2/CO ratio (Case
2). It can be concluded that the hybrid catalytic-dielectric barrier
discharge plasma reactor is potential for co-generation of synthesis
gas and higher hydrocarbons from methane and carbon dioxide and
performed better than the conventional fixed-bed reactor with respect
to CH4 conversion, C2+ yield and H-2 selectivity. (C) 2007 Published
by Elsevier Ltd.

2007







* 563(<-566): Hybrid artificial neural network-genetic algorithm technique for modeling and optimization of plasma reactor

A hybrid artificial neural network-genetic algorithm (ANN-GA)
numerical technique was successfully developed to model, to simulate,
and to optimize a dielectric barrier discharge (DBD) plasma reactor
without catalyst and heating. Effects of CH4/CO2 feed ratio, total
feed flow rate, and discharge voltage on the performance of
noncatalytic DBD plasma reactor were studied by an ANN-based
simulation with a good fitting. From the multiobjectives optimization,
the Pareto optimal solutions and corresponding optimal process
parameter ranges resulted for the noncatalytic DBD plasma reactor
owing to the optimization of three cases, i.e., CH4 conversion and C2+
selectivity, CH4 conversion and C2+ yield, and CH4 conversion and H-2
selectivity.

2006


* 564(<-156): Esterification of oleic acid to biodiesel using magnetic ionic liquid: Multi-objective optimization and kinetic study

The esterification of oleic acid in the presence of magnetic ionic liquid, 1-butyl-3-methylimidazolium tetrachloroferrate ([BMIM][FeCI4]) at reaction temperature of 65 degrees C has been investigated. Artificial neural network-genetic algorithm (ANN-GA) was used to simultaneously optimized methyl oleate yield and oleic acid conversion for the reaction. It was found that optimum responses for both yield and conversion were 83.4%, which can be achieved using molar ratio methanol-oleic acid of 22:1, catalyst loading of 0.003 mol and reaction time at 3.6 h. Esterification of oleic acid at optimum condition using recycled [BMIM][FeCl4] registered not much loss in catalytic activity after six successive runs. Kinetic study indicated that the reaction followed a pseudo-first order reaction, with activation energy and pre-activation energy of 17.97 kJ/mol and 181.62 min(-1), respectively. These values were relatively low compared to homogeneous or heterogeneous catalysts for esterification of oleic acid. Thus, [BMIM][FeCI4] is a promising new type of catalyst for conversion of high free fatty acid (FFA) feeds to biodiesel. (C) 2013 Elsevier Ltd. All rights reserved.

2014


* 565(<-179): Optimization of oleic acid esterification catalyzed by ionic liquid for green biodiesel synthesis

In order to improve the efficiency of biodiesel production from esterification of free fatty acids, an alternative to sulfuric acid has been explored. In this study, esterification of oleic acid was performed using 1-butyl-3-methylimidazolium hydrogen sulfate ([BMIM][HSO4]) ionic liquid for green biodiesel production. Response surface methodology (RSM) based on central composite design (CCD) was employed to study the effect of independent parameters on the process and also for single-objective optimization, while artificial neural network-genetic algorithm (ANN-GA) was utilized to simultaneously optimize the responses of the reaction (methyl oleate yield and oleic acid conversion). From the results, the predicted mathematical models for both methyl oleate yield and oleic acid conversion covered more than 80% of the variability in the experimental data. Furthermore, the linear temperature coefficient was identified as the most influential coefficient towards both responses. Higher responses were predicted for multi-objective optimization using ANN-GA compared to single-objective optimization using RSM. The optimum responses predicted using multi-objective optimization were 81.2% and 80.6% for methyl oleate yield and oleic acid conversion, respectively. The conditions to achieve optimum response were methanol-oleic acid molar ratio of 9:1, catalyst loading (0.06 mol), reaction temperature (87 degrees C), and reaction time (5.2 h). Furthermore, there was only small decrease in the catalytic activity of the IL after being recycled for five successive runs. (C) 2013 Elsevier Ltd. All rights reserved.

2013





* 575(<- 78): Modeling and optimization of catalytic performance of SAPO-34 nanocatalysts synthesized sonochemically using a new hybrid of non-dominated sorting genetic algorithm-II based artificial neural networks (NSGA-II-ANNs)

The effects of ultrasound-related variables on the catalytic
properties of sonochemically prepared SAPO-34 nanocatalysts in
methanol to olefins (MTO) reactions were investigated. Different
catalytic behaviors are observed which can be explained by the
differences in the catalysts' physicochemical properties affected by
ultrasonic (US) power intensity, sonication temperature, irradiation
time and sonotrode size. This result confirms that the activity of
SAPO-34 catalysts improves with the rise in US power, time and
temperature. In order to find a catalyst with the maximum conversion
of methanol, maximum light olefins content and maximum lifetime, a
hybrid of non-dominated sorting genetic algorithm-II based artificial
neural networks (NSGA-II-ANNs) was used. The multilayer feed forward
neural networks with back-propagation structures were implemented
using different training rules in the neural networks approach to
relate the ultrasound-related variables and the catalytic performance
of SAPO-34 catalysts. A comparison between experimental and artificial
neural network (ANN) values indicates that the ANN model with a 3-10-3
structure using the Bayesian regulation training rule has the best fit
and can be used as a fitness evaluation inside the non-dominated
sorting genetic algorithm-II (NSGA-II). Also, multiple linear
regression (MLR) was used to predict these objective functions. The
results indicate a poor fit for the objective functions with a low
coefficient of determination. This confirms that the ANN technique is
more effective than the traditional statistical-based prediction
models. Finally, this ANN model was linked to the NSGA-II and
Pareto-optimal solutions were determined by the NSGA-II.

2015



* 580(<-253): Multiobjective Optimization Design of Spinal Pedicle Screws Using Neural Networks and Genetic Algorithm: Mathematical Models and Mechanical Validation

Short-segment instrumentation for spine fractures is threatened by relatively high failure rates. Failure of the spinal pedicle screws including breakage and loosening may jeopardize the fixation integrity and lead to treatment failure. Two important design objectives, bending strength and pullout strength, may conflict with each other and warrant a multiobjective optimization study. In the present study using the three-dimensional finite element (FE) analytical results based on an L-25 orthogonal array, bending and pullout objective functions were developed by an artificial neural network (ANN) algorithm, and the trade-off solutions known as Pareto optima were explored by a genetic algorithm(GA). The results showed that the knee solutions of the Pareto fronts with both high bending and pullout strength ranged from 92% to 94% of their maxima, respectively. In mechanical validation, the results of mathematical analyses were closely related to those of experimental tests with a correlation coefficient of -0.91 for bending and 0.93 for pullout (P < 0.01 for both). The optimal design had significantly higher fatigue life (P < 0.01) and comparable pullout strength as compared with commercial screws. Multiobjective optimization study of spinal pedicle screws using the hybrid of ANN and GA could achieve an ideal with high bending and pullout performances simultaneously.

2013




* 590(<-279): A new weighted optimal combination of ANNs for catalyst design and reactor operation: Methane steam reforming studies

Catalyst design and evaluation is a multifactorial multiobjective
optimization problem and the absence of well-defined mechanistic
relationships between wide ranging input-output variables has
stimulated interest in the application of artificial neural network
for the analysis of the large body of empirical data
available. However, single ANN models generally have limited
predictive capability and insufficient to capture the broad range of
features inherent in the voluminous but dispersed data sources. In
this study, we have employed a Fibonacci approach to select optimal
number of neurons for the ANN architecture followed by a new weighted
optimal combination of statistically-derived candidate ANN models in a
multierror sense. Data from 200 cases for catalytic methane steam
reforming have been used to demonstrate the veracity and robustness of
the integrated ANN modeling technique. (C) 2011 American Institute of
Chemical Engineers AIChE J, 58: 24122427, 2012

2012




* 592(<-410): Multi-objective optimization of decoloration and lactosucrose recovery through artificial neural network and genetic algorithm

Artificial neural network (ANN) and genetic algorithm (GA) with
uniform design (UD) were used to optimize the decoloration and
lactosucrose (LS) recovery in solution with granular charcoal. Three
input variables (dosage of charcoal, time, and temperature) were
chosen in constructing the back propagation neural networks (BPNN)
model, and decoloration rate and LS recovery rate as output
variables. GA was used to optimize the input space of the ANN model to
find out the Pareto-optimal set. The best parameters were the dosage
of charcoal varying from 2.1894 to 2.1897%, time from 64.05 to 64.06
min, and temperature from 74.22 to 78.90 degrees C. The optimal
predicted decoloration rate is 96.30% and LS recovery rate is
97.35%. Results from confirmative studies showed that decoloration
rate was 94.85% and LS recovery rate was 97.23%, and the relative
error of network predicted values and actual measured values were
1.51% and 0.12%, respectively. The results suggested that the
UD-ANN-GA could effectively solve the separation efficiency by column
chromatography and the method was reliable.

2010



* 593(<-427): A Sequence Optimization Strategy for Chromatographic Separation in Reversed-Phase High-Performance Liquid Chromatography

A sequence optimization strategy combining an artificial neural
network (ANN) and a chromatographic response function (CRF) for
chromatographic separation it? reversed-phase high-performance liquid
chromatography has been proposed. Experiments were appropriately
designed to obtain unbiased data concerning the effects of varying the
mobile phase composition, flow-rate, and temperature. The ANN was then
used to simultaneously predict the resolution and analysis time, which
are the two most important features of chromatographic
separation. Subsequently, a CRF consisting of resolution and analysis
time was used to predict the optimum operating conditions for
different specialized purposes. The experimental chromatograms were
consistent with those predicted for given conditions, which verified
the applicability of the method. Furthermore, the proposed
optimization strategy was applied to literature data and very good
agreement was obtained. The results show that a strategy of sequential
combination of ANN and CRF can provide a more flexible and efficient
optimization method for chromatographic separation. (C) 2009 American
Institute of Chemical Engineers AIChE J, 56: 371-380, 2010

2010





* 596(<-242): Optimum culture medium composition for rhamnolipid production by pseudomonas aeruginosa AT10 using a novel multi-objective optimization method

BACKGROUND: Rhamnolipid is a biosurfactant that finds wide
applications in pharmaceuticals and beauty products. Pseudomonas
aeruginosa is a producer of rhamnolipids, and the process can be
implemented under laboratory-scale conditions. Rhamnolipid
concentration depends on medium composition namely, carbon source
concentration, nitrogen source concentration, phosphate content and
iron content. In this work, existing data7 were used to develop an
artificial neural network-based response surface model (ANN RSM) for
rhamnolipid production by pseudomonas aeruginosa AT10. This ANN RSM
model is integrated with non-dominated sorting differential evolution
(DE) to identify the optimum medium composition for this
process. RESULTS: Different strategies for optimization of culture
medium composition for this process were evaluated, and the best
determined to be an ANN model combined with DE involving a combination
of Naive and Slow and e-constrained techniques. The optimal culture
medium is determined to have carbon source concentration of 49.86 g
dm-3, nitrogen source concentration of 4.99 g dm-3, phosphate content
of 1.42 g dm-3, and iron content of 17.12 g dm-3. The maximum
rhamnolipid activity was found to be 18.07 g dm-3, which compares
favorably with that previously reported (18.66 g dm-3), and is in fact
closer to the experimentally determined value of 16.50 g
dm-3. CONCLUSION: This method has distinct advantages over methods
using statistical regression models, and can be used for optimization
of other multi-objective biosurfactant production processes. (c) 2012
Society of Chemical Industry

2013





* 597(<- 28): Optimization of controlled release nanoparticle formulation of verapamil hydrochloride using artificial neural networks with genetic algorithm and response surface methodology

This study was performed to optimize the formulation of polymer-lipid
hybrid nanoparticles (PLN) for the delivery of an ionic water-soluble
drug, verapamil hydrochloride (VRP) and to investigate the roles of
formulation factors. Modeling and optimization were conducted based on
a spherical central composite design. Three formulation factors, i.e.,
weight ratio of drug to lipid (X-1), and concentrations of Tween 80
(X-2) and Pluronic F68 (X-3), were chosen as independent
variables. Drug loading efficiency (Y-1) and mean particle size (Y-2)
of PLN were selected as dependent variables. The predictive
performance of artificial neural networks (ANN) and the response
surface methodology (RSM) were compared. As ANN was found to exhibit
better recognition and generalization capability over RSM,
multi-objective optimization of PLN was then conducted based upon the
validated ANN models and continuous genetic algorithms (GA). The
optimal PLN possess a high drug loading efficiency (92.4%, w/w) and a
small mean particle size (similar to 100 nm). The predicted response
variables matched well with the observed results. The three
formulation factors exhibited different effects on the properties of
PLN. ANN in coordination with continuous GA represent an effective and
efficient approach to optimize the PLN formulation of VRP with desired
properties. (C) 2015 Elsevier B.V. All rights reserved.

2015





* 599(<-390): Neural model for the leaching of celestite in sodium carbonate solution

A neural model for computing the conversion kinetics of SrSO(4) to
SrCO(3) was investigated in sodium carbonate solution, based on the
multilayered perceptrons was presented. For this purpose the
artificial neural network (ANN) method was used. The effects of
stirring speed, temperature, mole ratio Na(2)CO(3):SrSO(4) and
particle size of the celestite on leaching kinetics were studied. The
surface transformation of celestite to strontium carbonate in aqueous
carbonate solutions was also supported by FT-IR spectroscopy. The
conversion rate of celestite increases systematically with increasing
temperature (up to 70 degrees C). Furthermore, the feasibility of
replacing the SO(4)(2-) ions with CO(3)(2-) ions in the structure of
the leached solid was also investigated by FT-IR. FT-IR results showed
that the replacement of SO(4)(2-) ions in celestite by CO(3)(2-) ions
in leaching conditions was nearly completed at 60 degrees C with a
mole ratio Na(2)CO(3):SrSO(4) = 4:1, solid to liquid =5:500, -212+106
mu m particle size, and 400 rpm stirring rate for an interval of 240
min. The first (up to 90 min) conversion result obtained was trained
with an extended delta-bar-delta algorithm (EDBD), which is in the
multilayered perceptions and is a neural model structure. Results of
other conversion times (90-240 min) results were predicted. Results
predicted by the neural model were in very good agreement with the
experimental results. (C) 2010 Elsevier B.V. All rights reserved.

2010



* 601(<-376): Maximizing the native concentration and shelf life of protein: a multiobjective optimization to reduce aggregation

A multiobjective optimization was performed to maximize native protein
concentration and shelf life of ASD, using artificial neural network
(ANN) and genetic algorithm (GA). Optimum pH, storage temperature,
concentration of protein, and protein stabilizers (Glycerol, NaCl)
were determined satisfying the twin objective: maximum relative area
of the dimer peak (native state) after 48 h of storage, and maximum
shelf life. The relative area of the dimer peak, obtained from size
exclusion chromatography performed as per the central composite design
(CCD), and shelf life (obtained as turbidity change) served as
training targets for the ANN. The ANN was used to establish
mathematical relationship between the inputs and targets (from
CCD). GA was then used to optimize the above determinants of
aggregation, maximizing the twin objectives of the network. An almost
fourfold increase in shelf life (similar to 196 h) was observed at the
GA-predicted optimum (protein concentration: 6.49 mg/ml, storage
temperature: 20.8 degrees C, Glycerol: 10.02%, NaCl: 51.65 mM and pH:
8.2). Since no aggregation was observed at the optimum till 48 h, all
the protein was found at the dimer position with maximum relative area
(64.49). Predictions of the finally adapted network also reveal that
storage temperature and solvent glycerol concentration plays key role
in deciding the degree of ASD aggregation. This multiobjective
optimization strategy was also successfully applied in minimizing the
batch culture period and determining optimum combination of medium
components required for most economical production of actinomycin D.

2011





* 605(<- 32): Evaluation and prediction of membrane fouling in a submerged membrane bioreactor with simultaneous upward and downward aeration using artificial neural network-genetic algorithm

This paper describes the effect of simultaneous upward and downward
aeration on the membrane fouling and process performances of a
submerged membrane bioreactor. Trans-membrane pressure (TMP) and
membrane permeability (Perm) were simulated using multi-layer
perceptron and radial basis function artificial neural networks
(MLPANN and RBFANN). Genetic algorithm (GA) was utilized in order to
optimize the weights and thresholds of the models. The results
indicated that the simultaneous aeration does not significantly
improve the removal efficiency of contaminants. The removal
efficiencies of BOD, COD, total nitrogen, NH4+ - N and TSS were 97.5%,
97%, 94.6%, 96% and 98%, respectively. It was observed that the TMP
increases and the Perm decreases as operational time increases. The
TMP increasing rate (dTMP/dt) and the Perm decreasing rate (dPerm/dt)
for the upward aeration were 2.13 and 2.66 times higher than that of
simultaneous aeration, respectively. The training procedures of TMP
and Perm models were successful for both RBFANN and MLPANN. The train
and test models by MLPANN and RBFANN showed an almost perfect match
between the experimental and the simulated values of TMP and Perm. It
was illustrated that the GA-optimized ANN predicts TMP and
permeability more accurately than a network with a trial-and-error
approach calibration. (C) 2015 The Institution of Chemical
Engineers. Published by Elsevier BM. All rights reserved.

2015




* 606(<-338): Separation of toluene/n-heptane mixtures experimental, modeling and optimization

In this paper a composite membrane is used to separate toluene from
n-heptane mixture. The aim is to optimize the separation process
conditions through modeling. Therefore this model should be able to
predict membrane performance demonstrated by total permeation flux and
toluene selectivity as a function of operating condition. In order to
create a black box model of the process, a multi layer feed forward
artificial neural network is used. An algorithm based on evaluating
all possible structures is employed to create an optimum ANN
model. Number of hidden layers, transfer function, training method and
hidden neurons are determined with the aid of this
algorithm. Performance confirms that there is good agreement between
the experimental data and the model predicted values, with correlation
coefficients of more than 0.99 and mean square errors of less than
1%. Both model and experimental data show that increasing temperature
and toluene concentration increase total flux and decrease toluene
selectivity but increasing permeate pressure decreases both. Having
created and trained an optimized ANN model a multi-objective genetic
algorithm is employed to find optimum operating conditions with
respect to permeation flux and toluene selectivity as two targets of
this separation. Considering the obtained Pareto set and corresponding
decision variables, it is found that permeate pressure in this set is
almost constant and only variations in temperature and feed
concentration eventuate to the creation of the Pareto front. (C) 2011
Elsevier B.V. All rights reserved.

2011




* 608(<- 61): Modeling and optimization of antidepressant drug Fluoxetine removal in aqueous media by ozone/H2O2 process: Comparison of central composite design and artificial neural network approaches

Modeling and optimization of Fluoxetine degradation in aqueous
solution by ozone/H2O2 process was investigated using central
composite design (CCD) and the results were compared with the
artificial neural network (ANN) predicted values. We studied the
influence of basic operational parameters such as ozone concentration,
initial concentration of H2O2 and Fluoxetine and reaction time. The
ANN model was developed by feed-forward back propagation network with
trainscg algorithm and topology (4: 8: 1). A good agreement between
predicted values of Fluoxetine removal using CCD and ANN with
experimental results was observed (R-2 values were 0.989 and 0.975 for
the ANN and CCD models, respectively). The results showed that ANNs
were superior in capturing the nonlinear behavior of the system and
could estimate the values of Fluoxetine removal efficiency
accurately. Pareto analysis indicated that all selected factors and
some interactions were effective on removal efficiency. It was found
that the reaction time with a percentage effect of 45.04% was the most
effective parameter in the ozone/H2O2 process. The maximum removal
efficiency (86.14%) was achieved at ozone concentration of 30 mg L-1,
initial H2O2 concentration of 0.02 mM, reaction time of 20 min and
initial Fluoxetine concentration of 50 mg L-1 as the optimal
conditions. (C) 2014 Taiwan Institute of Chemical Engineers. Published
by Elsevier B.V. All rights reserved.

2015





* 612(<-623): Application of multivariate calibration and artificial neural networks to simultaneous kinetic-spectrophotometric determination of carbamate pesticides

A method for the simultaneous determination of the pesticides,
carbofuran, isoprocarb and propoxur in fruit and vegetable samples has
been investigated and developed. It is based on reaction kinetics and
spectrophotometry, and results are interpreted with the aid of
chemometrics. The analytical method relies on the differential rates
of coupling reactions between the hydrolysis product of each carbamate
and 4-aminophenol in the presence of potassium periodate in an
alkaline solution. The optimized method was successfully tested by
analyzing each of the carbamates independently, and linear calibration
models are described. For the simultaneous determinations of the
carbamates found in ternary mixtures, kinetic and spectral data were
processed either by three-way data unfolding method or decomposed by
trilinear modeling. Subsequently, 10 different RBF-ANN, PARAFAC and
NPLS calibration models were constructed with the use of synthetic
ternary mixtures of the three carbamates, and were validated with a
separate set of mixtures. The performance of the calibration models
was then ranked on the basis of several different figures of merit
with the aid of the multi-criteria decision making approach, PROMETHEE
and GAIA. RBF-ANN and PC-RBF-ANN were the best performing methods with
%Relative Prediction Errors (R-PE) in the 3-4% range and Recovery of
about 97%. When compared with other recent studies, it was also noted
that RBF-ANN has consistently outperformed the more common prediction
methods such as PLS and PCR as well as BP-ANN. The successful RBF-ANN
method was then applied for the determination of the three carbamate
pesticides in purchased vegetable and fruit samples. (C) 2004 Elsevier
B.V. All rights reserved.

2004


* 613(<-654): Spectrophotometric determination of metal ions in electroplating solutions in the presence of EDTA with the aid of multivariate calibration and artificial neural networks

Metal ions such as Co(II), Ni(II), Cu(II), Fe(III) and Cr(Ill), which
are commonly present in electroplating baths at high concentrations,
were analysed simultaneously by a spectrophotometric method modified
by the inclusion of the ethylenediaminetetraacetate (EDTA) solution as
a chromogenic reagent. The prediction of the metal ion concentrations
was facilitated by the use of an orthogonal array design to build a
calibration data set consisting of absorption spectra collected in the
370-760 nm range from solution mixtures containing the five metal ions
earlier. With the aid of this data set, calibration models were built
based on 10 different chemometrics methods such as classical least
squares (CLS), principal component regression (PCR), partial least
squares (PLS), artificial neural network's (ANN) and others. These
were tested with the use of a validation data set constructed from
synthetic solutions of the five metal ions. The analytical performance
of these chemometrics methods were characterized by relative
prediction errors and recoveries (%). On the basis of these results,
the computational methods were ranked according to their performances
using the multi-criteria decision making procedures preference ranking
organization method for enrichment evaluation (PROMETHEE) and
geometrical analysis for interactive aid (GAIA). PLS and PCR models
applied to the spectral data matrix that used the first derivative
pre-treatment were the preferred methods. They together with
ANN-radial basis function (RBF) and PLS were applied for analysis of
results from some typical industrial samples analysed by the
EDTA-spectrophotometric method described. DPLS. DPCR and the ANN-RBF
chemometrics methods performed particularly well especially when
compared with some target values provided by industry. (C) 2002
Elsevier Science B.V. All rights reserved.

2002




* 620(<-642): Neural network based optimization of drug formulations

A pharmaceutical formulation is composed of several formulation
factors and process variables. Several responses relating to the
effectiveness, usefulness, stability, as well as safety must be
optimized simultaneously. Consequently, expertise and experience are
required to design acceptable pharmaceutical formulations. A response
surface method (RSM) has widely been used for selecting acceptable
pharmaceutical formulations. However, prediction of pharmaceutical
responses based on the second-order polynomial equation commonly used
in an RSM, is often limited to low levels, resulting in poor
estimations of optimal formulations. The purpose of this review is to
describe the basic concept of the multi-objective simultaneous
optimization technique, in which an artificial neural network (ANN) is
incorporated. ANN's are being increasingly used in pharmaceutical
research to predict the nonlinear relationship between causal factors
and response variables. Superior function of the ANN approach was
demonstrated by the optimization for typical numerical examples. (C)
2003 Elsevier B.V. All rights reserved.

2003


* 621(<-668): Formula optimization based on artificial neural networks in transdermal drug delivery

The promoting effect of O-ethylmenthol (MET) on the percutaneous
absorption of ketoprofen from alcoholic hydrogels was evaluated in
rats in vitro and in vivo. Furthermore, a novel simultaneous
optimization technique incorporating an artificial neural network
(ANN) was applied to a design of a ketoprofen hydrogel containing
MET. When a small quantity of MET (0.25-0.5%) was added to the
hydrogels, the permeation of ketoprofen increased remarkably, compared
with the control. On the other hand, little change in permeation was
observed when small amounts of menthol were used (<1%), and at least
2% menthol was required to obtain a promoting efficiency comparable
with 0.25% MET. The partitioning of ketoprofen from the hydrogel to
the skin was improved by the addition of a small amount of MET,
whereas the diffusivity of the drug was enhanced at higher
concentration of MET (0.5-1%). For the optimization study, the amount
of ethanol and MET were selected as causal factors. A rate of
penetration (R-p) and lag time (t(L)) and total irritation score (TIS)
were selected as response variables. A set of causal factors and
response variables was used as tutorial data for ANN and fed into a
computer. Nonlinear relationships between the causal factors and the
response variables were represented well with the response surface
predicted by ANN. The optimization of the ketoprofen hydrogel was
performed according to the generalized distance function method. The
observed results of R-p and TIS, which had a lot of influence on the
effectiveness and safety, coincided well the predictions. (C) 1999
Elsevier Science B.V. All rights reserved.

1999


* 622(<-672): Artificial neural network as a novel method to optimize pharmaceutical formulations

One of the difficulties in the quantitative approach to designing pharmaceutical formulations is the difficulty in understanding the relationship between causal factors and individual pharmaceutical responses. Another difficulty is desirable formulation for one property is not always desirable for the other characteristics. This is called a multi-objective simultaneous optimization problem. A response surface method (RSM) has proven to be a useful approach for selecting pharmaceutical formulations. However, prediction of pharmaceutical responses based on the second-order polynomial equation commonly used in RSM, is often limited to low levels, resulting in poor estimations of optimal formulations. The aim of this review is to describe the basic concept of the multi-objective simultaneous optimization technique in which an artificial neural network (ANN) is incorporated. ANNs are being increasingly used in pharmaceutical research to predict the non-linear relationship between causal factors and response variables. The usefulness and reliability of this ANN approach is demonstrated by the optimization for ketoprofen hydrogel ointment as a typical numerical example, in comparison with the results obtained with a classical RSM approach.

1999



* 623(<-676): Multi-objective simultaneous optimization based on artificial neural network in a ketoprofen hydrogel formula containing O-ethylmenthol as a percutaneous absorption enhancer

The aim of this study was to apply a novel simultaneous optimization
technique incorporating an artificial neural network (ANN) to a design
of a ketoprofen hydrogel containing O-ethylmenthol (MET). For model
formulae, 12 kinds of ketoprofen hydrogels were prepared. The amount
of ethanol and MET were selected as causal factors. A percutaneous
absorption study in vivo in rats was performed and irritation evoked
on rat skin was microscopically judged after the end of the
experiments. The rate of penetration (R-p), lag time (t(L)) and total
irritation score (TIS) were selected as response variables. A set of
causal factors and response variables was used as tutorial data for
ANN and fed into a computer. Nonlinear relationships between the
causal factors and the release parameters were represented well with
the response surface predicted by ANN. The optimization of the
ketoprofen hydrogel was performed according to the generalized
distance function method. The observed results of R-p and TIS, which
had a lot of influence on the effectiveness and safety, coincided well
with the predictions. It was suggested that the multi-objective
simultaneous optimization technique incorporating ANN Was quite useful
for optimizing pharmaceutical formulae when pharmaceutical responses
were nonlinearly related to the formulae and process variables. (C)
1997 Elsevier Science B.V.

1997



* 624(<-677): Multi-objective simultaneous optimization technique based on an artificial neural network in sustained release formulations

The aim of this study was to develop a multi-objective simultaneous
optimization technique in which an artificial neural network (ANN) was
incorporated. As model formulations, 18 kinds of Trapidil tablet were
prepared. The amounts of microcrystalline cellulose, hydroxypropyl
methylcellulose and compression pressure were selected as causal
factors. In order to characterize the release profiles of Trapidil,
the release order and the rate constant were estimated. A set of
release parameters and causal factors was used as tutorial data for
ANN and fed into a computer. Non-linear relationships between causal
factors and the release parameters were represented well with the
response surface of ANN. The simultaneous optimization of the
sustained-release tablet containing Trapidil was performed by
minimizing the generalized distance between the predicted values of
each response and the optimized one that was obtained
individually. The optimal formulations gave satisfactory release
profiles, since the observed results coincided well with the predicted
results. These findings demonstrate that a multi-objective
optimization technique incorporating ANN is quite useful in the
optimization of pharmaceutical formulations. (C) 1997 Elsevier Science
B.V.

1997





* 626(<-153): Informatics based design of prosthetic Ti alloys

Informatics based approaches are employed to find a suitable
composition of Ti alloy, with high strength, low elastic modulus,
adequate biocompatibility and low cost. Artificial neural network,
capable of prediction and diagnosis in non-linear and complex systems,
is used to obtain the relationship of composition and processing
parameters with elastic modulus and yield strength. As the objectives
are conflicting, multiobjective optimisation using genetic algorithm
is employed to optimally design titanium alloys suitable for
prosthetic applications using the above models as objective functions
for the mechanical properties. The Pareto solutions provide the
desired alloy compositions where such properties may be achieved.

2014



* 632(<-421): Multi-target spectral moment QSAR versus ANN for antiparasitic drugs against different parasite species

There are many of pathogen parasite species with different
susceptibility pro. le to antiparasitic drugs. Unfortunately, almost
QSAR models predict the biological activity of drugs against only one
parasite species. Consequently, predicting the probability with which
a drug is active against different species with a single unify model
is a goal of the major importance. In so doing, we use Markov Chains
theory to calculate new multi-target spectral moments to fit a QSAR
model that predict by the first time a mt-QSAR model for 500 drugs
tested in the literature against 16 parasite species and other 207
drugs no tested in the literature using spectral moments. The data was
processed by linear discriminant analysis (LDA) classifying drugs as
active or non-active against the different tested parasite
species. The model correctly classifies 311 out of 358 active
compounds (86.9%) and 2328 out of 2577 non-active compounds (90.3%) in
training series. Overall training performance was 89.9%. Validation of
the model was carried out by means of external predicting series. In
these series the model classified correctly 157 out 190, 82.6% of
antiparasitic compounds and 1151 out of 1277 non-active compounds
(90.1%). Overall predictability performance was 89.2%. In addition we
developed four types of non Linear Artificial neural networks ( ANN)
and we compared with the mt-QSAR model. The improved ANN model had an
overall training performance was 87%. The present work report the
first attempts to calculate within a unify framework probabilities of
antiparasitic action of drugs against different parasite species based
on spectral moment analysis. (C) 2010 Elsevier Ltd. All rights
reserved.

2010




* 666(<-475): MODELING AND OPTIMIZATION OF A PHARMACEUTICAL FORMULATION SYSTEM USING RADIAL BASIS FUNCTION NETWORK

A Pharmaceutical formulation is composed of several formulation
factors and process variables. Quantitative model based pharmaceutical
formulation involves establishing mathematical relations between the
formulation variables and the resulting responses, and optimizing the
formulation conditions. In a formulation system involving several
objectives, the desirable formulation conditions for one property may
not always be desirable for other characteristics, thus leading to the
problem of conflicting objectives. Therefore, efficient modeling and
optimization techniques are needed to devise an optimal formulation
system. In this work, a novel method based on radial basis function
network (RBFN) is proposed for modeling and optimization of
pharmaceutical formulations involving several objectives. This method
has the advantage that it automatically configures the RBFN using a
hierarchically self organizing learning algorithm while establishing
the network parameters. This method is evaluated by using a trapidil
formulation system as a test bed and compared with that of a response
surface method (RSM) based on multiple regression. The simulation
results demonstrate the better performance of the proposed RBFN method
for modeling and optimization of pharmaceutical formulations over the
regression based RSM technique.

2009