MULTI-LEVEL DIMENSIONALITY REDUCTION METHODS USING FEATURE SELECTION AND FEATURE EXTRACTION

This paper presents a novel feature selection method called Feature Quality (FQ) measure based on the quality measure of individual features. We also propose novel combinations of twolevel and multi level dimensionality reduction methods which are based on the feature selection like mutualcorrelation, FQ measure and feature extraction methods like PCA(Principal ComponentAnalysis)/LPP(Locality Preserving Projection). These multi level dimensionality reduction methodsintegrate feature selection and feature extraction methods to improve the classification performance. In theproposed combined approach, in level 1 of dimensionality reduction, feature are selected based on mutualcorrelation and in level 2 features are selected from the reduced set obtained from level 1 based on thequality measure of the individual features and vice versa. In another proposed combined approach, featureextraction methods like PCA and LPP are applied on the reduced feature set obtained in level 1 or /andlevel 2. To evaluate the performance of the proposed methods several experiments are conducted onstandard datasets and the results obtained show superiority of the proposed methods over single leveldimensionality reduction techniques