中文摘要：

为了可以实时推荐符合人们情感状态的音乐，提出了一种融入音乐子人格特质的社交网络行为分析的音乐推荐算法，该算法通过分析用户发表在微博等社交媒体上的状态，计算用户在该情感状态下对音乐的偏好程度：选择在该情感状态下音乐偏好相似的最近邻用户，最后融入音乐子人格特质进行偏好度计算．为用户推荐最适合其情感状态的音乐。实验结果表明，该算法可以缓解用户数据稀疏性对推荐结果的影响，能够提高推荐系统的推荐质量。

英文摘要：

In order to recommend music conforming to the people＇s sentiment state in real time, a novel music recommendation method combining music sub-personality and social network behavior analysis was proposed. Through sentiment analysis of the users＇ words and sentences which was released on Weibo and other social network, the method calculated the users＇ music preference in a sentiment state, the most suitable music was recommend to users in that sentiment state. Experimental results show that the proposed method can alleviate the effect of user data sparsity on the recommendation results.