Feature Selection on High Dimensional Data Using Wrapper Based Subset Selection

Abstract

In recent years, feature subset selection and classification in high dimensional data is a major challenge faced by the researchers. The main aim of the feature subset selection is to find most informative features from the vast number of features in the high dimensional data. Filter, wrapper and embedded methods are currently used to solve these issues. In this paper, we have incorporated wrapper based subset selection technique for selecting a subset from the high dimensional datasets. In this approach to find the optimal threshold value, the feature subsets are given to the classifier iteratively until the maximum accuracy is obtained. The symmetrical uncertainty method is used to weight the features to predict the predominant feature. For validating the incorporated algorithm, we have used 10-fold cross validation against the two standard classification techniques such as Naive Bayes and Support Vector Machine (SVM) and the results are tabulated and compared. The comparison between the results shows that the proposed method gives the better accuracy and results.