An Enhanced Entropy-K-Nearest Neighbor Algorithm Based on Attribute Reduction

Dimension disaster will directly impact on the efficiency and accuracy of K-nearest neighbor (KNN) classification algorithm. In order to reduce the effect, this chapter proposes an improved Entropy-KNN algorithm based on attribute reduction. It combines KNN algorithm with information entropy theory to reduce the attribute, and the test sample is classified by the average distance and the numbers on the respective class. The experimental results show that compared with traditional KNN algorithm, the proposed algorithm enormously raises the classification accuracy rate; meanwhile it also maintains the efficiency of category.