论文部分内容阅读
为了获取用户的信息需求,并依据信息需求模型在因特网上搜索相关文本,文章提出了基于示例的用户信息需求模型的获取和表示方法。其基本思想是:在用户给定的示例文本集的基础上,利用特征项的类别区分度,抽取能够表现用户兴趣的项作为用户信息需求模型的基本特征项集。然后,基于统计上的Fisher准则,进行判别分析,以获取特征项在相关文本的判定中的重要程度。最后,给出用户信息需求模型的逻辑和物理表示。
In order to obtain the information needs of users and to search related texts on the Internet according to the information requirement model, the article proposes an example-based method for obtaining and presenting user information requirement models. The basic idea is: based on the set of example texts given by the user, using the category distinguishing degree of the feature items, the items that can express the user’s interest are extracted as the basic feature item sets of the user information requirement model. Then, based on the statistical Fisher criterion, discriminant analysis is conducted to obtain the significance of feature items in the judgment of related texts. Finally, the logical and physical representation of the user information needs model is given.