A New Filter Approach Based on Generalized Data Field

In this paper, a new feature selection method based on generalized data field(FR-GDF) is proposed. The goal of feature selection is selecting useful features and simultaneously excluding garbage features from a given feature set. It is Important to measure the "distance" between data points in existing feature selection approaches. To measure the "distance", FR-GDF adopts potential value of data field. Information entropy of potential value is used to measure the inter-class distance and intra-class distance. This method eliminates unimportant or noise features of original feature sets and extracts the optional features. Experiments prove that FR-GDF algorithm performs well and is independent of the specific classification algorithm.