K-means Clustering Algorithm Based on Distance Threshold and Weighted Sample

An improved K-means clustering algorithm is proposed based on distance threshold and weighted sample. First the sample mean of sample set is selected as the first initial clustering center; secondly clustering center and clustering number are dynamically determined based on distance threshold; finally the method of weighted sample to reduce the influence of the clustering effect, the weighted sample points participate in the whole process of clustering and the clustering quality of different clustering algorithm are measured based on silhouettete coefficient. The experimental results show that, compared with the original K-means text clustering algorithm and the algorithm in reference1, the proposed algorithm can improve the clustering quality.