一种改进蚁群聚类算法在数据挖掘中的研究  

An Improved Ant Colony Clustering Algorithm in Data Mining

作　　者：侯超[1] 侯永[1] 

机构地区：[1]上海理工大学光电与计算机工程学院计算机系,上海200093

出　　处：《微计算机信息》2011年第12期136-138,共3页Control & Automation

摘　　要：聚类分析是数据挖掘的一个重要分支,是认识和探索数据之间内在联系的有效方法。以往的蚁群聚类算法中蚂蚁只能作为个体在空间内的移动式随机的,这样的随机移动算法时间复杂度比较大,而且效率也不高。本论文研究了一种基于信息素控制蚂蚁移动的策略。通过蚂蚁间接通讯进行交流和学习。智能个体蚂蚁通过在环境中留下的信息素来完成环境的修改,而其他的蚂蚁则通过间接通讯策略来对环境做出相应的反应,从而降低了算法的时间复杂度。Cluster analysis is an important branch of data mining and it is a effective method to recognize the intrinsic link between the data. In the previous ant colony clustering algorithm ,the ant only as individuals in the space of mobile random, so that the random movement algorithm time complexity is relatively large and inefficient.This study provide a pheromone-based control strategy of ants moving.Ants communicate indirectly through communication and learning. Intelligent individual ants by leaving the information in the environment and modify the environment. The other ants through the indirect communication strategies to respond appropriately to the environment,thereby reducing the time complexity of the algorithm.

关 键 词：聚类分析 蚁群算法 信息素 间接通讯 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]