摘要
为了解决个性化网站中很少考虑用户检索意图,检索效果较差的问题,提出了一种有效的增量协同过滤推荐方法。该增量协同过滤推荐模型改进了最流行的推荐算法之一的协同过滤算法,并应用到个性化网站中。通过分析Web日志提取用户的浏览行为,将其归一化为用户对项目的评分值,并利用改进的相似度计算方法得到用户之间的相似度值,从中选择能够表现用户偏好的最近邻集合进行评分预测后对结果排序,将排序后的结果作为推荐列表返回给用户。最后设计增量更新算法实时有效地更新用户的历史偏好数据。实验表明,增量协同过滤推荐模型适用于个性化网站,利用该方法可以使推荐结果更加符合用户意图。
In order to solve the problems that user’s retrieval intention is rarely considered in the personalized websites and the search result is not good, this paper proposes an Incremental Collaborative Filtering Recommendation method(ICFR). The ICFR model improves the collaborative filtering recommendation algorithm, which is one of the most popular recommendation algorithms, and applies it to the personalized websites. Firstly, the browsing behavior information of users is extracted by analyzing web logs and normalized into rating value. Secondly, the similarity value between users is obtained by using the improved similarity calculation method, and the nearest neighbor set which can reflect the user’s intention is selected. Thirdly, the results predicted by the nearest neighbor set are sorted and returned to the user as the recommended list. Finally, historical user preference data are updated effectively in real time by the incremental updating algorithm. The experimental results indicate that the incremental collaborative filtering recommendation model is suitable for personalized websites and this method can make the retrieval results reflect the user intention.
作者
李婷
张瑞芳
郭克华
LI Ting;ZHANG Ruifang;GUO Kehua(School of Information Science&Engineering,Central South University,Changsha 410083,China;Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education,Nanjing University of Science and Technology,Nanjing 210094,China)
出处
《计算机工程与应用》
CSCD
北大核心
2019年第4期225-232,共8页
Computer Engineering and Applications
基金
国家自然科学基金(No.61202341)
高维信息智能感知与系统教育部重点实验室创新基金(No.JYB201502)
科技部国家国际科技合作专项项目(No.2013DFB10070)
湖南省创新平台专项(No.2012GK4106)
中南大学中央高校基本科研业务费专项资金(No.2017zzts718)
关键词
个性化网站
基于用户的协同过滤算法
推荐系统
用户意图
增量式更新
personalized website
user-based collaborative filtering
recommendation system
user intention
incremental updating