摘要
基于蚁群算法的聚类算法已经在当前的数据挖掘研究中得到应用。针对蚁群聚类算法早期出现的缺点,提出一种蚁群聚类组合方法使其得以改进。改进思路是引入K-means作为蚁群算法的预处理过程。通过K-means快速、粗略地确定聚类中心,利用K-means方法的结果作为初值,再进行蚁群算法聚类。有效地解决了蚁群算法早期收敛过慢等问题。
The ant-based clustering algorithm is applicated in the data mining community.Due to the disadvantage of the classical algorithm,this paper presents an improved ant colony clustering combination method.The paper introduces K-means to take the ant colony algorithm the pre-eomputation process.Through K-means,it definites cluster center lastly and sketchily,and takes the starting value using the K-means method result,again executes the ant colony algorithm cluster. It solves the ant colony algorithm for early slow convergence effectively.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第18期146-148,共3页
Computer Engineering and Applications
基金
国家科技支撑计划项目No.2007BAH08B04~~
关键词
聚类
蚁群算法
信息素
聚类组合
clustering
ant colony algorithm
pheromone
clustering
