Accuracy comparison between gene selection methods using NAIVE Bayes classifier for the microarray data of JEV infected Mus Musculus brain cells

Japanese Encephalitis is the most important cause of epidemic encephalitis worldwide. From the reports by various sources, about 68,000 cases of Japanese encephalitis (JE) are estimated to occur each year [1]. A vaccine is available for Japanese encephalitis, which utilizes effectively killed inoculated bacteria, but it is expensive and requires a primary vaccination followed by two successive boosters. No successful cure is discovered till date. This paper describes the analysis of microarray data of Japanese Encephalitis Virus (JEV) infected Mus Musculus cells performing computational accuracy comparison between four gene selection methods using a classifier. The accuracy check is utilized for the effective selection of differentially expressed genes which can be a possible drug target for JE. The proposed methodology includes the comparison of gene selection methods; Logfold selection, ttest, Forward feature selection using fuzzy entropy and Sequential forward selection with the help of the Naive Bayes classifier. The result derived shown that the Forward Feature Selection method using Fuzzy entropy proves to be the maximum accurate gene selection method using the Naive Bayes classifier and in the case of this particular data used.