Genetic Algorithm with Feature Selection Technique for Nominal Data

Abstract

Data mining methods are used to handle the problems of dynamic huge data set. We can scale back the time complexity of calculated by selecting only useful features to build a classification model. A feature selection technique is used to select only useful features from available features. We have used an intersection principle based feature selection approach. In genetic algorithm is used as a search method and in this we select only the features which are appears frequently in datasets. Then we tested the result on different datasets having different type of data using Naive Bayes & J48 classifiers. The result analysis shows that Naí¯ve Bayes classifier gives better result than J48 classifier, with the substantial growth in accuracy, minimum time and minimum number of features. We have used Correlation feature selection with Genetic Algorithm for feature selection.