Comparing Performance of Recommendation Techniques in the Blogsphere
msra(2006)
摘要
Weblogs are one of fundamental components of Web 2.0 and there are a lot of unskilled bloggers and visitors who have difficulties in finding relevant blogs. Recommender systems are a so- lution to the information overload problems. In this paper a weblog recommender system based on link structure of weblog graph is in- troduced. Here we treat links between weblogs as some kind of rat- ing, meaning that if there is a hyperlink from weblog A pointing to weblog B then blogger A prefers weblog B. Based on this assumption we have adapted almost all of famous recommendation techniques for Blogsphere. The methods are implemented on a real dataset and experiments show promising results.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络