时间加权的混合推荐算法

发布时间：2018-05-16 02:26

本文选题：个性化推荐 + 时间权重　；参考：《计算机科学》2016年S2期

【摘要】：提出了一种基于时间加权的混合推荐方法。该方法分为离线和在线两个阶段,离线阶段根据目标用户对物品的评价等信息得到与目标用户有相似兴趣的邻居,并构建物品描述模型;在线阶段根据目标用户和邻居用户的评价行为构建用户描述模型。由于用户兴趣会随外部因素而产生概念漂移,因此在算法中引入衰减系数以提高推荐质量。在滑动窗口模型下,每隔一定时间间隔,更新用户模型和相似群组,产生个性化的推荐。实验结果表明,该算法能实时反映用户兴趣,提高推荐系统的准确率,有较高的用户满意度。
[Abstract]:A hybrid recommendation method based on time weighting is proposed. The method is divided into two stages: offline and online. In the off-line phase, the neighbors with similar interest to the target user are obtained according to the information such as the target user's evaluation of the item, and the article description model is constructed. In the online stage, the user description model is constructed according to the evaluation behavior of the target user and the neighbor user. Due to the concept drift of user interest with external factors, the attenuation coefficient is introduced to improve the quality of recommendation. In the sliding window model, user models and similar groups are updated at regular intervals to generate personalized recommendations. The experimental results show that the proposed algorithm can reflect the interest of users in real time and improve the accuracy of recommendation system.
【作者单位】：金陵科技学院信息化建设与管理中心;扬州大学信息学院计算机系;南京大学软件新技术国家重点实验室;金陵科技学院计算机工程学院;
【基金】：国家自然科学基金项目(61379066) 江苏省高校自然科学基金项目(15KJD520008) 江苏省现代教育技术研究课题(2014-R-32521)资助
【分类号】：TP391.3

【相似文献】