网络信息过滤中的固定文章集表达方法

为解决在信息过滤（ｉｎｆｏｒｍ ａｔｉｏｎ ｆｉｌｔｅｒｉｎｇ）中表达用户的信息需求困难的问题，提出了一种新方法——固定文章集法（ｆｉｘｅｄ ｄｏｃｕｍ ｅｎｔｓｅｔ， Ｆ Ｄ Ｓ）。在这种方法中，用户可以通过评价一些专门选择的文章来表达自己的信息需求。这避免了用户选择关键词的困难，扩大了用户信息需求表达的空间，同时可以应用于处理多语种信息过滤和社会化过滤问题。选择固定文章集使用选择文章基因法（ｓｅｌｅｃｔｄｏｃｕｍ ｅｎｔ＇ｓｇｅｎｅ， Ｓ Ｄ Ｇ），实验证明， Ｓ Ｄ Ｇ 的性能较其它方法有明显优势

FDS expressive method in information filtering

A novel method, fixed document set (FDS) method, was proposed regarding solving the difficulty of represent user's information need in information filtering. According to the FDS, user's information need is represented by carefully selected documents from a particular area together with user's evaluations. This could reduce difficulties on selecting key words, expand user's expressive band. The method could also be used in multilingual retrieval system and social filtering system. SDG (select document's gene) algorithm was used for selecting FDS. The experiment proved that the performance of SDG is better than that of the other selecting methods.