Research of liver cancer detection based on improved K-NN algorithm

Jianhua Liu, Jianwei Wang and

Abstract

As a simple and effective classification algorithm, k-Nearest Neighbor (K-NN) algorithm is widely used in many fields. In order to improve the efficiency of classification in Liver Cancer Detection, the Principal Component Analysis (PCA) method is applied to the K-NN algorithm, which selects the effective features efficiently. Building new attribute sets and applying new effective features to K-NN classification separately, to obtain the correct classification rates of new effective features. Then the correct classification rates are applied to the K-NN Algorithm for classification as the distance weight. The improved K-NN Algorithm has been applied to Liver Cancer Detection, and the experiment indicated has obtained the good effect.