A Novel Approach Feature Selection based on Neighborhood Positive Region (NPR)

Due to increase in large number of document on the internet data mining becomes an important key parameter. Numerous data mining techniques are being carried for extracting the valuable information such as clustering, classification and cluster analysis. In the field of machine learning, pattern recognition and data mining, feature selection also called as attribute reduction becomes a challenging problem. Also the key lies in reducing the attributes and selecting the relevant features. Hence, to overcome the issues of attribute reduction we proposed Neighborhood positive region (NPR) based on rough set theory. In this paper we have shown the experimental result of NPR is implemented on three UCI data sets which show the computational time and reduced features