【作者】 杜亚军； 裴峥； 彭宏； 徐扬；

【机构】 西南交通大学智能控制中心； 西华大学计算机科学与工程系；

【摘要】 <正>近年来,搜索引擎的个性化是一个研究热门话题,在文[1]中王实等提出一种新的基于分类方法的实时个性化推荐方法:该文首先根据用户访问事务文法生成序列访问事务集,用于得到每个用户访问的序列特性,并用次序列输入分类器进行分类,以得到用户的下面一些可能访问的页面,为用户提供实更多还原

【Abstract】 Humanity method of topic recommendation in search engine is put forward.It this paper,we analyze the interest of person in same interest group exists a certain similarity,and so the same interest group that they together concern about same topics is identified through clustering analysis.In the process of searching Web information, we exercise the uncertain reasoning in light of the regular grammar.It is composing the URL recommend sequence by the regular grammar that the Interviewing topic of user is it’s head.The URLs of relevant interest topic of the another user is the other nodes.It manifest sufficiently that the regular grammar method is very suitable to recommend URLs to user by experiment’s analysis.更多还原