Artificial intelligence

Abstract: A new approach to solving the problem of joint constructing of classification decision rules and feature selection in the recognition training systems based on weighted training samples use is given in the article. W-GridDC method of forming a weighted training sample of w-objects, the algorithm w-MIEF of construction the feature set based on it, and a modified method of k-nearest neighbor for objects classification are described. Experimental results have confirmed the effectiveness of the proposed approach.

Keywords: adaptive recognition system, training samples, w-object, feature selection

References: