基于分类和聚类相结合的个性化检索方法研究

Research on personalized search based on classification and clustering

目前检索工具的设计大多面向所有用户,而没有考虑到用户个人的兴趣,导致查准率较低。由此提出一种基于分类和聚类相结合的个性化信息检索方法。该方法首先利用聚类技术,对用户的历史浏览记录进行聚类,产生一个初步的用户兴趣分类,然后根据ODP对用户兴趣分类进行调整得到最终的用户兴趣分类,并利用该分类对传统搜索引擎返回的结果进行分类,以产生有意义的分类搜索结果。该方法克服了单独利用分类或聚类技术的局限性,提高了搜索引擎的可用性。

At present, the design of information search tools were mostly based on the needs of all users, not based on the personal interest, and thus lead to lower precision. A kind of personalized search method based on classification and clustering technology is introduced in this paper. First of all, elementary user interest classes are obtained through the clustering the pages that has been visited, and then the classes are adjusted referencing ODP web directory to get the final interest classes which later will be used for search results sorting. The localization which using only classification or clustering technology is gotten over. The usability of searching engine is enhanced.