Feature Subset Selection Algorithm for High Dimensional Data using Fast Clustering Method

Abstract

Feature selection means finding most useful features and it will produce suitable results among entire set of features. An algorithm is used to selecting a feature and it may be evaluated from both efficiency and effectiveness point of view. Efficiency is related to the time required to find a subset of features while the effectiveness is related to quality of subset of features. Based on these, we proposed a fast clustering-based feature selection algorithm (FAST). FAST algorithm performs in two steps. First of all, features are divided into various clusters. Then the most useful feature is selected from each cluster. We adopt the minimum spanning tree (MST) to increase the efficiency of FAST. Many useful feature selection algorithms such as FCBF, Relief, CFS, Consist, FOCUS-SF are compared to FAST algorithm.

Cite this paper

@inproceedings{GowriDurga2014FeatureSS, title={Feature Subset Selection Algorithm for High Dimensional Data using Fast Clustering Method}, author={A . GowriDurga and A . Gowri Priya}, year={2014} }