摘要
HITS是Web结构挖掘代表算法之一.HITS算法仅考虑页面之间的链接关系,容易出现主题偏移现象.文中分析了用户历史行为与兴趣之间的关系,从四个方面近似计算用户兴趣度.提出了基于用户兴趣度的改进HITS算法(HC-HITS),对比实验结果表明HC-HITS算法减少了主题漂移现象,取得较好的效果.
HITS is one of the representative algorithms of Web structure mining. Because HITS algorithm only considers the relations among pages, it is easy to present the subject deviation phenomenon. Based on analyzing the relation between user historical behaviors and interest,the paper calculates approximately user interest from 4 aspects. After that the paper presents improved HITS algorithm （HC-HITS） based on user interest degree. Through comparison, the experiment results show HC-HITS algorithm reduces subject devi- ation phenomenon and obtains good effect.
出处
《南华大学学报（自然科学版）》
2013年第2期54-57,共4页
Journal of University of South China：Science and Technology
基金
2011年湖南省教育厅科研基金资助项目(11C1097)
2012年南华大学高等教育研究与改革立项基金资助项目(2012XJG15)
