An Efficient Clustering-Classification Method in an Information Gain NRGA-KNN Algorithm for Feature Selection of Micro Array Data

Gene expressions by microarray data technique have been effectively utilized for classification and diagnostic of cancer nodules. Numerous data mining techniques like clustering are presently applied for identifying cancer using gene expression data. An unsupervised learning technique is a clustering technique used to find out grouping structure in a set of data. The problem of feature selection in clustering algorithm is what type of data attributes used is not known and also for data there is no class labels so there is no clear criteria to direct the search. A further issue in clustering is the identification of the number of clusters that affects the performance of feature selection. Gene expression database have a great potential as a medical diagnostic tool since they represent the state of a cell at the molecular level. Training data sets is available for the classification of cancer types generally have a fairly small sample size compared to the number of genes involved. Feature selection is considered to be a problem of optimization in machine learning, reduces the number of features, noisy and redundant data, and results in acceptable classification accuracy. Hence, selecting significant genes from the microarray data poses a dreadful challenge to researchers due to their high-dimensionality features in clustering technique and the usually small sample size. A clustering algorithm is proposed, which is a hybrid model of information gain genetic algorithm for feature selection in microarray data sets. Information Gain (IG) was used to select important feature subsets (genes) from all features in the gene expression data, and a Non-Dominated Ranked Genetic Algorithm (NRGA) was employed for actual feature selection. The K-NN method is used to evaluate the NRGA algorithm. Experimental results show that the proposed clustering based method simplifies the number of gene expression levels effectively and gives accurate feature selection while compared with other methods. [Akey Sungheetha, Dr. J.Suganthi. An Efficient Clustering-Classification Method in an Information Gain NRGA-KNN Algorithm for Feature Selection of Micro Array Data. Life Sci J 2013; 10(7s): 691-700] (ISSN:1097-8135). http://www.lifesciencesite.com. 108 Keywords---Feature Selection, Gene Expression, Genetic Algorithm, Non-Dominated Ranked Genetic Algorithm, Information Gain, K-nearest neighbor (K-NN)