Performance issues in genetic Information retrieval

Abstract

Information retrieval IR systems are responsible for retrieval of large newlineamounts of data in an efficient manner The information retrieval task deals with newlineretrieving data stored within the repositories in response to a user s information newlineneeds For this type of retrieval system a performance system is necessary This newlineresearch explores the data mining techniques in order to identify the one that will newlineoffer the best performance in application to bioinformatics that will respond to newlinegenetic information Data mining has been exploited to retrieve the valuable newlineinformation in a wide spread fields especially in DNA microarray technology newlineThe DNA microarray technology produces a huge amount of gene data newlineexpression levels of thousands of genes for a very few samples newlineInitially performance issues for information retrieval in the field of newlinemicroarray gene data are analyzed The information retrieval metrics like accuracy newlinesensitivity specificity are to be increased and error rate is to be reduced As a first newlineprocess in gene classification the high dimensionality of the microarray gene data newlineis reduced using LPP The LPP is chosen for the dimensionality reduction because newlineof its ability of preserving locality of neighborhood relationship Secondly SVM newlinehas been trained for effectual gene classification SVM has the ability to learn with newlinevery few samples and so it is selected SVM ensemble was utilized for the newlineclassification process in order to classify more than a single class newline