| 孙克雷,陈安东.基于用户兴趣的个性化推荐算法研究[J].安徽建筑大学学报,2017,25(1):65-69 |
| 基于用户兴趣的个性化推荐算法研究 |
| Research on Personalized Recommendation Algorithm Based on User Interest |
| |
| DOI：10.11921/j.issn.2095-8382.20170115 |
| 中文关键词: 用户兴趣 协同过滤 时间窗 个性化推荐 |
| 英文关键词: User interest collaborative filtering time window personalized recommendation |
| 基金项目:安徽省自然科学基金(1408085QE94) |
|
| 摘要点击次数: 8192 |
| 全文下载次数: 0 |
| 中文摘要: |
| 针对协同过滤算法存在用户兴趣不易捕捉的问题，提出了一种基于用户兴趣偏移和项目自身属性特征的个性化推荐算法。利用滑动时间窗内项目属性和用户评分建立出用户兴趣偏爱因子，通过推荐项目自身属性特征给出用户对项目的偏爱度；最后结合项目偏爱度和协同过滤算法中预测评分产生推荐。实验结果表明，该算法准确反映出用户兴趣的偏移和项目自身属性特征，在推荐质量上也得到提高。 |
| 英文摘要: |
| Aiming at the problem that user's interest is not easy to capture by the collaborative filtering algorithm, a personalized recommendation algorithm based on the changes of users’ interest and the self-characteristic of items is proposed. The interest preference factors of users are established by using items attributes and user rating within the sliding time windows. Then the items preference degrees of user are given by the characteristic of recommended items themselves. Finally, the recommendation is produced according to the preference degree of items and the predictive score in the collaborative filtering algorithm. Experimental results show that the proposed algorithm can accurately reflect the changes of users’ interest and the attribute of items themselves. In the meanwhile, the quality of recommendation is improved compared to the classical UserCF method. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
