【摘要】 利用Web数据挖掘技术,对用户未来的访问进行预测和推荐,是实现网站柔性个性化服务的研究方向之一.笔者通过对网站个性化服务相关技术的研究,改进了对Web服务器用户访问日志信息进行聚类分析的关联数据竞争聚类(competitiveagglomerationforrelationdata,CARD)算法,并在此基础上构建了相应的实验模型.CARD算法是CA算法的一种改进算法,适合于处理没有明显特征的Web数据.试验是建立在真实的Web日志上的,实验证明该算法具有较好的聚类效果和适用度.更多还原

【Abstract】 Using Web data mining technology to predict the user’s future request and make recommendation is one aspect of personalized web service. A clustering algorithm CARD is presented to analyze user access logs of Web server and an experimental model was constructed on the basis of analysis. CARD(Competitive Agglomeration for Relation Data) is based on CA that can adaptively process Web data. The tests are based on realistic Web logs. CARD has proved to be more applicable than CA in Web data mining and yielded superior results.更多还原