Animesh Kumar Paul

Genetic algorithm based fuzzy decision support system for the diagnosis of heart disease

Decision in medical diagnosis is mostly taken by experts experiences. In many cases, not all the experts experiences contribute towards effective diagnosis of a disease. Researchers have taken multiple approaches like attribute reduction, rule extraction, fuzzy model optimization, etc. But noisy data in datasets, irrelevant attributes, and lack of effective fuzzy rules are major hindrances to provide best decision. In this study, we propose genetic algorithm based fuzzy decision support system for predicting the risk level of heart disease. Our proposed fuzzy decision support system (FDSS) works as follows: i) Preprocess the dataset, ii) Effective attributes are selected based on different methods, iii) Weighted fuzzy rules are generated on the basis of selected attributes using GA, iv) Build the FDSS from the generated fuzzy knowledge base, v) Predict the heart disease. The experiments carried out with real-life data set show the effectiveness of this proposed innovative approach.

New Automatic Fuzzy Relational Clustering Algorithms Using Multi-Objective NSGA-II

Abstract Most of the fuzzy clustering methods use only centroid information and cannot differentiate the geometric structures of clusters due to the cohesion and separation measures of the fuzzy partition. Moreover, conventional clustering methods confine the search space of automatic fuzzy clustering as it shows a tendency to fall into local minima as well as the number of clusters is required to be provided as prior knowledge. In this paper, combining fuzzy relational clustering (FRC) with multi-objective genetic algorithm NSGA-II, we propose two new fuzzy relational clustering methods, referred to as FRC NSGA and IFRC NSGA. In IFRC NSGA, NSGA-II is used to adjust the parameters of FRC algorithm dynamically such as the appropriate number of clusters and the initial membership values. Thus this proposed model handles the intricacy of automatic clustering a dataset, at the same time, avoids the problem of falling into local minima quickly. For both of the proposed models, NSGA-II is used to optimize two cluster validity indices named, separation, and cohesion, simultaneously, whereas, FRC is used to handle the overlapping properties of clusters. In this case, fuzzy membership degrees are used to compute the overlap-separation. To eliminate the impediment of automatic fuzzy clustering, it is necessary to optimize one or more criteria. Therefore, we consider NSGA-II as an optimization algorithm to optimize more than single criterion. The experimental results have exhibited that the intended methods provide competitive results in the compound, overlapped, high-dimensional gene expression and non-gene expression datasets. The proposed methods are able to determine well-separated, hyperspherical, noncompact and overlapping clusters compared to other existing methods.

Adaptive weighted fuzzy rule-based system for the risk level assessment of heart disease

Expert’s knowledge base systems are not effective as a decision-making aid for physicians in providing accurate diagnosis and treatment of heart diseases due to vagueness in information and impreciseness and uncertainty in decision making. For this reason, automatic diagnostic fuzzy systems are very time demanding to improve the diagnostic accuracy. In this paper, we have developed an automatic fuzzy diagnostic system based on genetic algorithm (GA) and a modified dynamic multi-swarm particle swarm optimization (MDMS-PSO) for prognosticating the risk level of heart disease. Our proposed fuzzy diagnostic system (FS) works as follows: i) Preprocess the data sets ii) Effective attributes are selected through statistical methods such as Correlation coefficient, R-Squared and Weighted Least Squared (WLS) method, iii) Weighted fuzzy rules are formed on the basis of selected attributes using GA, iv) MDMS-PSO is employed for the optimization of membership functions (MFs) of FS, v) Build the ensemble FS from the generated fuzzy knowledge base by fusing the different local FSs. Finally, to ascertain the efficiency of the adaptive FS, the applicability of the FS is appraised with quantitative, qualitative and comparative analysis on the publicly available different real-life data sets. From the empirical analysis, we see that this hybrid model can manage the knowledge vagueness and decision-making uncertainty precisely and it has yielded better accuracy on the different publicly available heart disease data sets than other existing methods so that it justifies its adaptability with different data sets.

Incorporating gene ontology into fuzzy relational clustering of microarray gene expression data

The product of gene expression works together in the cell for each living organism in order to achieve different biological processes. Many proteins are involved in different roles depending on the environment of the organism for the functioning of the cell. In this paper, we propose gene ontology (GO) annotations based semi-supervised clustering algorithm called GO fuzzy relational clustering (GO-FRC) where one gene is allowed to be assigned to multiple clusters which are the most biologically relevant behavior of genes. In the clustering process, GO-FRC utilizes useful biological knowledge which is available in the form of a gene ontology, as a prior knowledge along with the gene expression data. The prior knowledge helps to improve the coherence of the groups concerning the knowledge field. The proposed GO-FRC has been tested on the two yeast (Saccharomyces cerevisiae) expression profiles datasets (Eisen and Dream5 yeast datasets) and compared with other state-of-the-art clustering algorithms. Experimental results imply that GO-FRC is able to produce more biologically relevant clusters with the use of the small amount of GO annotations.

Reconstruction of gene network through Backward Elimination based Information-Theoretic Inference with Maximal Information Coefficient

For understanding the complex processes of regulation within the system of cellular and every process of life in different developmental and environmental contexts, reconstructing Gene Regulatory Networks(GRNs) is an essential part of Systems Biology. A recently developed maximal information coefficient (MIC) is better to detect all kinds of association than others and it maintains both generality and equitability properties. In this study, we combined MIC as an entropy estimator with gene regulatory network method Backward Elimination based Information-Theoretic Inference and then compare this proposed method with the MI-based algorithm MRNETB by examining SynTReNs datasets. The performance of our proposed MIC based MRNETB (MRNETB-MIC) is given by using both the receiver operator characteristic (ROC) curve and the precision-recall (PR) curve and from these, the proposed method shows significantly better performance in reconstructing gene regulatory network.

Optimizing fuzzy membership function using dynamic multi swarm — PSO

Performance of fuzzy application to solve the control problems depends on a number of parameters such as the choice and shape of the membership function. Defining MFs manually in a proper way is time consuming, prone to errors and difficult. And especially it depends subjectively based on experts experiences. Improvement of the performance of the fuzzy control system is made by the optimization of the membership function. In this paper, a Dynamic Multi-Swarm PSO is used to optimize the fuzzy membership function. DMS-PSO has the ability to avoid local optimal and able to generate an optimal set of parameters for fuzzy control system. The experiment carried out with real-life application, park a vehicle into garage beginning from any starting position; results show that the better performance of proposed fuzzy model is obtained by using the optimized membership functions than a simple fuzzy model when the membership functions were heuristically defined.

A multi-objective genetic algorithm based fuzzy relational clustering for automatic microarray cancer data clustering

In the biological and biomedical research field, Microarray technology has grown into a leading approach. Microarray technology helps to monitor a large number of genes simultaneously based on different experimental conditions. This paper propose a fast elitist non-dominated sorting multi-objective genetic algorithm (NSGA-II) based fuzzy relational clustering approach for clustering microarray cancer expression dataset where it smartly creates a finer trade-off between fuzzy compactness and fuzzy separation of the clusters. Binary encoding schema is used for chromosomes while encoding the multi-objective GA. In this case it encodes the variable length numbers of clusters. Fuzzy Relational Clustering method is used to automatically assign the membership to all data point for each cluster and it produce a clusters for the given microarray cancer dataset. In the multi-objective genetic algorithm, a set of non-dominated solutions is given with optimizing objectives without dominating to any other solution. The simulation results exhibits that the intended method gives promising results in the complex, overlapped, high-dimensional microarray cancer datasets.

Optimization of interval type-2 fuzzy logic controllers with rule base size reduction using genetic algorithms

This paper presents optimization technique to develop type-2 fuzzy systems (FSs) through hybrid genetic algorithms (HGAs). The proposed optimization technique works as follows: (i) Optimize the type-2 membership functions (ii) Learn the rule base through genetic algorithms (iii) Apply the reducing technique to reduce the rule base. (iv) Build the FSs based on type-2 membership functions and the reduced rule base. For concurrently works step (i) and (ii), we used real and binary coded coupled GAs for the optimization technique. Real coded GAs is used to tune the type-2 membership functions and binary coded GAs is used to learn and reducing the fuzzy rules. For intelligent control of a two degree freedom inverted pendulum system, the control algorithm is used. Finally, the simulation studies show that the generated controller performance is better and comparable to the existing methods under normal conditions.

Stacked auto encoder training incorporating printed text data for handwritten bangla numeral recognition

Recognition of handwritten numerals has gained much interest in recent years due to its various application potentials. Bangla is a major language in Indian subcontinent and is the first language of Bangladesh; but unfortunately, study regarding handwritten Bangla numeral recognition (HBNR) is very few with respect to other major languages such as English, Roman etc. Some noteworthy research works have been conducted for recognition of Bangla handwritten numeral using artificial neural network (ANN) as ANN and its various updated models are found to be efficient for classification task. The aim of this study is to develop a better HBNR system and hence investigated deep architecture of stacked auto encoder (SAE) incorporating printed text (SAEPT) method. SAE is a variant of neural networks (NNs) and is applied efficiently for hierarchical feature extraction from its input. The proposed SAEPT contains the encoding of handwritten numeral into printed form in the course of pre-training and finally initializing a multi-layer perceptron (MLP) using these pre-trained weights. Unlike other methods, it does not employ any feature extraction technique. Benchmark dataset with 22000 hand written numerals with different shapes, sizes and variations are used in this study. The proposed method is shown to outperform other prominent existing methods achieving satisfactory recognition accuracy.

Sentiment mining from Bangla data using mutual information

Due to the explosion of social networking sites, blogs and review sites (for example, Amazon, Twitter, and Facebook, etc.) it provides an overwhelming amount of textual information. We need to organize, explore, analyze the information for making a better decision from the side of customers and companies. Thus, sentiment analysis is the best way in which it determines the authors feelings expressed in reviews as positive or negative opinions by analyzing an enormous number of documents. In this work, we used Mutual Information (MI) for the feature selection process and also used Multinomial Naive Bayes (MNB) for the classification of Bangla and English reviews. The experimental results demonstrate that the system can achieve satisfactory accuracy for Bangla dataset compare to English dataset where Bangla dataset is generated from Amazons Watches English dataset.