基于序列模式的用户浏览行为提取与分析

当今互联网所提供的功能和服务越来越多，Web内容也越来越丰富，移动应用越来越流行。然而，复杂的Web服务应用对用户提出了更高的要求，给用户浏览带来了很多问题，很多时候用户会感到无所适从。文中提出基于用户浏览序列模式的用户行为提取与分析方法。该方法可以分为浏览模式分析和用户聚类两部分。在浏览模式分析时，首先根据用户行为数据得到浏览序列，然后运用序列模式挖掘PrefixSpan算法获取用户习惯的浏览模式，最后把分析获取的用户浏览模式应用到Web浏览中，为不同的用户需求提供个性化的服务。在用户聚类时，运用层次聚类方法按照浏览模式的相似性对用户进行聚类，以分析用户的不同属性（如年龄、职业、学历等）对用户浏览模式的影响。实验结果表明，文中采用的PrefixSpan算法和层次聚类方法在用户浏览模式分析和研究方面具有很好的可行性和有效性。

User Browsing Behavior Extraction and Analysis Based on Sequence Pattern

Nowadays, intemet provides more and more application and services, there are more and more contents on Web, and mobile applications develop more and more quickly. Meanwhile, the more sophisticated Web services applications need higher quality of users. Many times some users may feel confused in those Web services. It presents the sequence based on user access patterns for user behavior extraction and analysis methods. The method can be divided into two parts of the browse mode analysis and user clustering. For browse mode analysis, first based on user behavior data get navigation sequence;Then the sequential pattern mining algorithm, so-called PrefixSpan, is used in getting user browsing pattern ; Finally applies user browsing pattern to Web services, and provides personalized services to different users. For user clustering, a hierarchical clustering method is used for splitting user into different categories. The impact of differ- ent attributes of user I such as age, occupation, education ） to user clustering is researched. Experimental results show that the PrefixSpan algorithm and hierarchical clustering method used in the browse mode analysis has a good feasibility and validity in this paper.