论文部分内容阅读
当前基于协作过滤(CF,Collaborative Filtering)的推荐系统广泛应用于在线购物、音乐点播、智能Web推荐等系统。基于协作过滤的Web推荐系统的一个问题是用户通常仅仅访问很少Web页,因此根据用户访问Web页的记录找到一组相似用户的概率很低,这就是“稀疏问题”。本文提出了一种利用WWW冲浪模型,模仿用户访问Web页过程中的一些特点,并将用户的冲浪过程延续,模拟用户在Web站点访问更多的Web页,从而估计出用户对更多Web页的评价。本文还给出了实验比较,表明扩展冲浪深度后,系统推荐Web页的效果得到明显提高。
The current collaborative filtering (CF, Collaborative Filtering) based recommendation system is widely used in online shopping, music on demand, intelligent Web recommendation system. One of the problems with collaborative filtering-based Web recommender systems is that users typically only have very few Web pages, so there is a very low probability of finding a similar set of users based on the Web page’s record of the user, which is called Sparse Problems. This paper presents a WWW surfing model to simulate the user access to the Web page in the process of some of the characteristics of the user’s surfing process and continue to simulate the user visits the Web site more Web pages to estimate the user for more Web pages evaluation of. This paper also gives an experimental comparison, which shows that after the surfing depth is extended, the effect of system recommendation Web page is obviously improved.