一种基于改进混合蛙跳的聚类算法

摘要

聚类分析是一种无监督的模式识别方式,它是数据挖掘中的重要技术之一.给出了一种基于改进混合蛙跳算法的聚类分析方法,该方法结合了K-均值算法和改进混合蛙跳算法各自的优点,引入了K-均值操作,再用改进混合蛙跳算法进行优化,很大程度上提高了该算法的局部搜索能力和收敛速度.通过仿真对基于改进混合蛙跳的聚类方法与其他已有的聚类方法进行了比较,验证了所提出算法的优越性.%Clustering analysis is an unsupervised mode of pattern recognition and is one of primary techniques in the filed of data mining. A clustering analysis method based on a modified shuffled frog leaping algorithm (MSFLA)is proposed. The new approach integrates the advantages of the MSFLA and the K-means algorithms, which introduces the K-means operation and utilizes the MSFLA for optimization, improves the locally searching capability and convergence speed of the clustering algorithm based on MSFLA. Simulations are performed to compare the performance of the clustering algorithm based on modified MSFLA and other clustering algorithm, which validates the effectiveness of the proposed algorithm.