Constrained Based Feature Subset Selection Algorithm for High Dimensional Data

IJSRD - International Journal for Scientific Research & Development| Vol. 2, Issue 10, 2014 | ISSN (online): 2321-0613

IJSRD Found Good research work on Computer Science & Engineering area...

Author(s):

G.Geethanjali , K.S.R. College of Engineering; Mr.P.Prakash, K.S.R. College of Engineering

Keywords:

Feature Selection, AR Relevancy, Redundancy, Entropy, Conditional Entropy

Abstract:

Feature Selection is to selecting the useful features from the original dataset for improve the more accurate results. Constrained Based Feature Subset Selection(CFSS) Algorithm Removes irrelevant and redundant features. This method is to find a similarity computation based on the entropy and conditional entropy values. After computing similarity computation to applied Approximate Relevancy(AR) algorithm which will find the relevance between the attribute and class labels from that computation most relevant attributes will be selected, then using Adaptive k++ neighborhood algorithm group those relevant features and create graph according to that relevant features. After calculating relevant features to form the spanning tree using kruskal’s algorithm, removing all redundant features for which it has an edge in tree.Finally, to select best subset of the features from the original dataset.

I. INTRODUCTION

Feature selection has been an active research area in pattern reorganization, statistics and data mining communication. Nowadays a rapid growth of high dimensional data such as digital images, gene expression microarrays, dimensionality reduction has been a fundamental tool for many data mining

tasks.

For Full Length article Click Here