摘要
作为数据挖掘技术的重要组成部分,聚类分析在很多领域有着广泛的应用。蚁群算法由于采用分布式并行处理和正反馈机制,具有较好的全局收敛性,并且在解决多种NP难问题中取得了成功。将信息素扩散模型引入到蚁群聚类算法中,通过设计新的信息素更新机制,提出一种新的基于信息素扩散的蚁群聚类算法。实验结果表明新算法在聚类效果上比基本的蚁群聚类算法有较明显的改善。
As an important part of data mining technology, clustering is widely used in many fields. By using distributed parallel computing and positive feedback mechanism, ant colony algorithms have optimal global convergence, and have succeeded in solving many NP-hard problems. A model of pheromone diffusion is introduced to the ant colony clustering algorithm and by using a new designed pheromone update mechanism, a novel ant colony clus- tering algorithm based on pheromone diffusion mechanism is proposed. The results of experiments show that the new method is improved obviously compare with ant colony clustering algorithm.
出处
《科学技术与工程》
2009年第16期4657-4661,共5页
Science Technology and Engineering
基金
国家自然科学基金(60874089)资助
关键词
聚类
蚁群算法
信息素扩散
蚁群聚类算法
clustering ant colony algorithm pheromone diffusion ant colony clustering algorithm