Feature selection is an important task in data mining and machine learning. Selecting relevant information from the Bio-medical data is a challenging task particularly microarray data in cancer research. In this paper, we propose a new feature selection algorithm based on Relief and Information gain concept called Relevance Gain-R (RG-R), the primary objective of the algorithm is to select optimal number of features from microarray dataset. The proposed algorithm calculates the relevant information between two instances (i.e. randomly selected instance) with respect to the target class, then it evaluates the neighbor distance between hit ‘H’ and miss ‘M’ owing by ReliefF. In this paper, we tested our algorithm with ten benchmarking microarray datasets and gained good results. However, for empirical analysis, we compared our approach with three existing feature selection algorithms, and also, we tested with three best classifiers i.e. classifier-1 C4.5, KNN, Naïve Bayes. The result exhibits that the proposed feature selection approach is efficient and very effective in selecting relevant features.
Volume 11 | 08-Special Issue
Pages: 2407-2418