基于层次概念的用户兴趣模型研究

Study of user profile technology based on hierarchy

使用基于关键词匹配的方法,分析了HTML语言描述的Web文档,提取网页中有用的特征信息,得到两类标记中的内容:一类是网页的全局描述信息,如和;另一类起局部修饰作用,强调了网页的部分内容,如。从而提出了基于层次概念的用户模型,并使用向量空间模型方法建立了以突发事件新闻为基础的用户兴趣模型。实验表明,这种方法有一定的可行性。

Web documents described by HTML language is analyzed by using the method about matches based on the key word, which the useful characteristic information of web documents. Thus two kinds marks contents is obtained： One is overall description information about web documents like 〈TITLE〉〈/TITLE〉and 〈META〉〈/META〉. Another played embellishment in web documents, emphasized partial contents more or less, like 〈Hn〉.thus proposed based on the hierarchy concept user model. Then, based web accidental news corpora, user interest model is established by vector space model. The experiment indicated that, the method has the certain feasibility.