基于内容和协同过滤相融合的推荐算法
2018-02-03沈华理
沈华理
摘要:协同过濾算法有两个主要问题:新用户冷启动问题和相似用户的可靠性问题。为了解决上述问题,提出了基于内容和协同过滤相融合的推荐算法,主要解决新用户冷启动、相似用户可靠性问题。该算法的主要过程为,利用k-means聚类算法将数据集中的用户进行聚类,然后确定用户各个属性特征的适当权重,根据用户人口统计学特征的聚类方法,将新用户分配到恰当的类中,最后提取出新用户的最近邻,根据最近邻用户的项目评分,计算新用户对未评分项目的预评分,生成推荐列表。实验结果表明,在平均绝对误差(MAE)和均方根误差(RMAE)上有较明显的改善。
关键词:协同过滤;冷启动;人口统计学特征;k-means聚类;混合推荐
中图分类号:TP391 文献标识码:A 文章编号:1009-3044(2018)02-0232-03
Recommendation Algorithm Based on the Combination of Content and Collaborative Filtering
SHEN Hua-li
(College of Computer Science and Engineering, Anhui University of Science and Technology, Huainan 232001, China)
Abstract: There are two main problems in collaborative filtering algorithm:the problem of new user cold start and the reliability of similar users. In order to solve the above problems, Recommendation Algorithm Based on the Combination of Content and Collaborative Filtering is proposed, which mainly solves the problem of cold start and similar user reliability. The main process of the algorithm is that the k-means clustering algorithm is used to cluster the users of the data set, then the appropriate weight of each attribute of the user is determined, and the new user is assigned to the appropriate class according to the clustering method of the user demographic characteristics, finally, the nearest neighbor of the new user is extracted. According to the project score of the nearest neighbor, Calculating the pre rating of a new user on a non rated project and generating a list of recommendations. The experimental results show that there is a significant improvement in the evaluation standard of MAE and RMSE.
Key words: collaborative filtering; cold start; demographic characteristics; k-means clustering; mixed recommended
推荐系统会在某些情况下被使用,像电影商店、图书馆、餐饮、旅游以及其他方面为用户提供有趣的选择和项目。特别在电子商务和网络在线电影系统中得到了广泛的应用。从过去到现在一直广受欢迎研究之一的推挤系统是电影推荐系统。在信息量巨大的情况下,在恰当的时间向用户提供最有吸引力的项目是个性化推荐领域[1]的相关内容之一。在电影推荐系统中,使得用户可以根据电影标题、导演、编剧和发布日期等特征寻找电影。总体来看,推荐系统被划分为两种主要的种类[2]:基于内容的推荐和协同过滤推荐。在基于内容的推荐算法[3]中,是根据用户分配给内容、新闻文本和链接的权重进行推荐,也就是把权重最高的项目向用户推荐。……
