Social tags have been acknowledged as a highly useful resource in retrieving music by moods or topics. However, since social tags are open for labeling, some social tags are inaccurate. In this paper, we present a new framework to identify accurate social tags of songs. In our framework, we first clean and filter music tags. Then we apply an improved hierarchical clustering algorithm to group the tags to build a tag category. Based on the category, we classify music songs using lyrics. In order to extend the semantic information of lyrics, we apply CLOPE to cluster lyrics and use the centroid of the corresponding cluster to represent the lyrics. Based on the Na\"ive Bayes method, the probability of assigning lyrics to particular class is predicted. The classification result is then used to determine whether a social tag is accurate. The experimental results show that the proposed framework is effective and encouraging.

We proposed a method to classify songs in the Million Song Dataset according to song genre. Since songs have several data types, we trained sub-classifiers by different types of data. These sub-classifiers are combined using both classifier authority and classification confidence for a particular instance. In the experiments, the combined classifier surpasses all of these sub-classifiers and the SVM classifier using concatenated vectors from all data types. Finally, the genre labels for the Million Song Dataset are provided.

With the rapid growth of web data, people sometime need semantic similar information in order to obtain a clear outline of their interests, so recommendation is needed to provide relevant information to users' queries. In this paper, we propose a method to recommend semantic similar movies and stars to users' queries, styles and stories. The system measures the similarities between movies according to genre and style features extracted from YAGO and IMDB. Experimental results show that the recommendations meet users' interests.