摘要
会话聚类是一种重要的Web使用挖掘技术,旨在发现相似的用户行为,这是目前电子商务中的热点问题之一。该问题的难度在于要对大规模的会话进行聚类,这些会话被表示成高维向量,加大了对算法高效性的要求。提出了一种ACO和PSO相结合的算法进行会话聚类分析。实验结果表明该算法与ACO算法、PSO算法、K-means算法相比,具有更好的性能。
Session clustering is an important technology of Web usage mining, aiming to find similar user behavior, which is one of the hot fields in electronic business. The difficulty of the problem lies in the large scale session data, which needs to be represented as the high dimensional vector, and which is a challenge to the performance of the algorithm. This paper presents a type of clustering algorithm combining ACO with PSO algorithm. Experimental results show that the algorithm has better perfor- mance compared with ACO, PSO and K-Means algorithm.
出处
《计算机工程与应用》
CSCD
2013年第22期136-138,218,共4页
Computer Engineering and Applications
基金
国家自然科学基金(No.N0.71071047)
安徽省自然科学基金项目(No.1208085MG120)
高等学校博士学科点专项科研基金(No.20090111110016)
合肥工业大学博士学位专项资助基金(No.2010HGBZ0301)
关键词
WEB使用挖掘
蚁群优化
粒子群优化
会话聚类
电子商务
Web usage mining
Ant Colony Optimization （ACO）
Partical Swarm Optimization （PSO）
session clustering
e-business
