关键词: 雷达信号预分选, K-Means算法, 聚类, 孤立点

Abstract:

K-Means algorithm finds wide applications in radar signal pre-sorting.The traditional K-Means algorithm depends largely on the number of clusters and the initial setting of cluster centers.It is also very sensitive to noise and outliners.In order to solve these problems,a radar signal pre-sorting method which combines the distance method with the improved K-Means algorithm is presented in this paper.Simulation shows that the proposed method effectively reduces the influence of noise and outliners on K-Means clustering algorithm.

Key words: radar signal pre sorting;K Means;clustering;outliers