基于标签扩展的协同过滤算法在音乐推荐中的应用

The Role of Tag-Extension Collaborative Filtering Algorithm in Music Recommendation

推荐技术作为过滤海量信息的手段,在音乐领域也给人们带来了便利。但音乐推荐与传统的电商推荐相比存在显式反馈不宜获取、内容特征提取代价大等缺陷。针对上述情况,提出一种基于标签扩展的协同过滤算法,将社会化标签内容作为物品内容,基于内容计算对用户未收听的物品进行评分,在此基础上利用协同过滤为用户提供推荐列表。实验结果表明,该算法可以有效改善推荐结果的准确性,提高推荐质量。

Recommended technology is a means of filtering large amounts of information, bringing convenience to people in the music field. But music recommendation is different from traditional electricity recommendation. There exist drawbacks in the music field, like difficult acquiring explicit feedback, costly extracting content features and so on. In view of the above situation, a collaborative filtering algorithm based on label extension is proposed. The content of socialized label is taken as the content of the item, and the score of the non-listened item is calculated based on the content. The user will get a recommended list by this way. The experimental results show that the algorithm can improve the accuracy of recommendation results and the recommended quality.