A new multiclass embedded feature selection method using genetic algorithm

In this paper, we propose an embedded subset selection method based on minimum redundancy–maximum relevance criterion, which uses Pierson's correlation coefficient criterion in redundancy and accuracy of nearest neighbor classification in relevancy. In this method first some features with low sensitivity are eliminated then remainder of original feature subset is used in subset selection process which uses genetic algorithm. Sensitivity of features shows correlation of each feature with target. The proposed method is tested over several well-known benchmarking datasets. The performance of the proposed algorithm is also compared with some recent hybrid filter–wrapper algorithms. The results show that this method is competitive in terms of both classification accuracy and the number of selected features.

A new multiclass embedded feature selection method using genetic algorithm Keywords:

component , Feature Selection , Embedded , Genetic Algorithm , Subset Selection , Sensitivity