论文部分内容阅读
为了根据用户的信息需求,在因特网上搜索相关文本,该文提出了一种文本过滤的匹配机制,其基本思想是:利用基于词典的概念扩张方法,改进用户模板。计算扩张的用户模板与文本的全局相似度,获取初步的过滤结果;在文本特征区域,进行标题、摘要段、首段和尾段等片断的局部相似度计算,以综合评价文本与用户模板的匹配情况。该方法可操作性强,效果明显。
In order to search for relevant texts on the Internet according to the information needs of users, this paper proposes a matching mechanism of text filtering. The basic idea is to improve the user template by using the concept expansion method based on the dictionary. Calculate the global similarity between the expanded user template and the text, and obtain the preliminary filtering result; in the text feature area, calculate the local similarity of the segments such as the title, the abstract segment, the first segment and the tail segment so as to comprehensively evaluate the similarity between the text and the user template Matching situation. The method is easy to operate and the effect is obvious.