Study on comprehensive evaluation model of attribute coordinate based on evaluation sample selection by K-means

摘要

It is a very important step that the sample points are marked by experts in the comprehensive evaluation method based on attribute coordinate. At present, the sample points are selected randomly in some algorithms. However, this method possibly causes that the sample points has the homogeneity and can not represent the whole sample space, thus affecting the precision of evaluation results. In this paper, K-means clustering method is used to select sample points for evaluation, and the corresponding simulation experiment is also carried out. Then the simulation results showed the advantages of improved algorithm.