CONSISTENCY EVALUATION FOR EXPONENTIAL, SEQUENTIAL AND RANDOM SEARCH STRATEGIES USING FILTER APPROACH

Document Type : Primary Research paper

Author

Department of Computer Science and Applications The Gandhigram Rural Institute Gandhigram, Dindigul, India

Abstract

Feature selection can improve the accuracy and efficiency of the learning process. Some of the methods are based on the search of the features that allow the data set considered consistent. Many alternative analysis functions that are employed in feature selection can be categorized as measures like distance, information, dependence, consistency, and classifier error rate. This research work is proposed on consistency measures by using exponential, sequential and random based searching strategies. The thought behind these measures is to predict the concept or class value of its instances. The consistencies that will study and compare all methods are dealt with in this paper elaborately.

Keywords