期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2017
卷号：6
期号：11
页码：21648
DOI：10.15680/IJIRSET.2017.0611095
出版社：S&S Publications
摘要：Data mining is the computing process of discovering patterns in large data sets involving methods at theintersection of machine learning, statistics, and database systems. Irrelevant, noisy and high dimensional data containlarge number of features, which degrades the performance of data mining and machine learning tasks. One of themethods used to reduce the dimensionality of data is feature selection. Feature selection method selects a subset offeatures that represents original features in problem area with high accuracy. Various methods have been proposed tofind these subsets. These methods either time consuming to find subset or support with optimality [1]. This paperpresents a new feature selection approach that combines Firefly Optimization algorithm (FFO) with RoughSetTheory.The algorithm suggests that the attraction system of fireflies shows the feature selection procedure. The experimentalresults show the comparative analysis of various measures such as accuracy, specificity, and sensitivity andcomputation time with different protein data sets namely Protein tolerance datasets, Protein stability datasets, mRNAsplice site datasets and Transcription factor binding site dataset
