基于RSS数据源的用户兴趣模型改进及应用

Improvement and application of a user interest model based on RSS data sources

在基于RSS数据源的用户兴趣模型研究领域,已有的模型是一种由用户模型根节点、信息类别和用户兴趣子类构成的三层结构的树状用户兴趣模型。本文改进了原有模型的表示形式以及关键词权重判定函数,将关键词权值计算方法中加入日期参数,并且将用户的兴趣分为长期兴趣和短期兴趣,针对两种兴趣采用不同的更新机制。实验表明,改进模型的个性化程度更高,对用户兴趣更新更及时准确。模型应用于农业信息领域,有助于农民及时获得自己需要的个性化农业信息,对农业信息化建设具有一定的应用价值。

A user interest model based on RSS data sources, with a tree structure which has three layers--user point, information sort, and user interest classes--has previously been proposed. In this paper, an improved form of the original model is described： The determination of keyword weight is improved by adding the date to it, and user＇s interest is divided into long-term interest and short-term interest, which have different update mechanisms. The experimental results indicate that the improved user interest model is well-personalized and efficiently updated. The model can be applied to help farmers obtain timely information and contribute to the development of agricultural information.