摘要
一个好的推荐算法在如今的智能Web应用中变得十分重要。给出一种具有时间认知的推荐算法。传统的推荐算法通常不加选择地使用早期和近期的评价,随着时间的推移而忽略了用户兴趣的变化,且用户的兴趣在短期的时间间隔内保持稳定但在长期间隔内是有所改变的。在已有的协同过滤推荐算法基础上加入了时间向量因子用来激励近期记录或减弱早期记录,以达到更有效分类相似兴趣的用户的目的。结果表明应用该方法能有效提高在智能Web中推荐的准确率及效率。
A good recommendation algorithm is becoming very important in today＇s intelligent web applications. In this paper,we propose a time cognitive recommendation algorithm. Traditional recommendation algorithms often use early and recent evaluations indiscriminately,ignoring user interest changes over time. And user interests remain stable over short periods of time but change over a long period of time. Based on the existing cooperative filtering recommendation algorithm,the time vector factor is used to motivate the recent records or weaken the early records in order to achieve the purpose of more effective classification of similar interests. The results show that the proposed method can effectively improve the accuracy and efficiency in the intelligent web.
出处
《计算机应用与软件》
2017年第2期58-63,共6页
Computer Applications and Software
基金
国家自然科学基金项目(61170060)
安徽省学术与技术带头人学术科研活动资助项目(2015D046)
安徽省高等学校优秀拔尖人才项目(gxbjZ D2016044)
关键词
智能WEB
时间向量因子
协同过滤推荐
用户兴趣变化
Intelligent web
Time vector factor
Collaborative filtering recommendation
User interest changes
