Enhanced gene expression classification for efficient diagnosis with optimized feature selection

Abstract

As the dimensionality of the data increases in machine learning, the amount of data needed to provide a reliable analysis increases exponentially. Microarray or gene expression profiling is applied to compare and determine the gene expression level and pattern for different cell types or tissue samples in a single experiment. The recent arrival of DNA microarray technology has led to the concurrent monitoring of thousands of gene expressions in a single chip which stimulates the progress in cancer classification. The grouping of different tumors of the gene expression data is very critical in the diagnosis of cancer and the discovery of the drug and is even more possible owing to the importance in the diagnosis of cancer owing to its huge size. These DNA based micro array technologies have resulted in expressing many thousands of genes in one single experiment and for the purpose of analysing the profiles of expression. The feature selection-based methods will choose the informative genes before the classification of the data of microarray for the prediction and diagnosis of cancer. These methods remove the redundant and irrelevant features for improving the accuracy of classification. In this work, proposed the Artificial Bee Colony (ABC) based feature selection in bone marrow PC gene expression data. Swarm intelligence-based ABC algorithm has been proposed to find the best features in the gene identification. The ABC will be used for selection that generates the subset of the features and used for every feature produced by the onlookers; therefore, this proposed system will be based on the lines of wrapper-based feature selection. The main goal to this is choosing the minimum number of genes which are deemed to be very significant for that of the PCs having the improvement of the accuracy of prediction by using this proposed approach. The results have shown that this method of ABC based feature selection for that of the GDS531 has a higher accuracy of classification with an accuracy of about 2.94% compared to the GDS2643. newline