SYSTEM AND METHOD FOR RECOMMENDING MUSIC COMBINING LISTENING HABIT AND TAG INFORMATION

摘要

The present invention relates to a music recommendation system and a method in a social music site using music listening habits of users and tag information. The music recommendation system of the present invention comprises: a user information collection unit for collecting user information including details about music that a user listens to and tag information applied to each music item; a user similarity extracting unit for extracting a group having user similarity in listening to similar music using the user information; and a music item recommendation unit for recommending music items using information about preference of music items of the group having user similarity. The user similarity extracting unit includes: a listening habit collection unit for collecting listening habits in listening to the music items of a user using user information; a tag information processing unit for collecting tag information applied to each music item using the user information; and a user similarity selection unit for selecting a group having user similarity in listening to music items similar to the music items of another user using the listening habits and tag information. Also, the tag information processing unit includes: a tag pre-processing unit for pre-processing the tag information in a standardized form; a tag categorizing unit for categorizing the pre-processed tag information into a fact tag and emotion tag; a weighted value assigning unit for assigning weighted value for the categorized tag information; a tag ontology interlocked with the tag pre-processing unit, storing information including homophones, abbreviation, loanword, and synonym for each natural language; and an emotion ontology for storing the weighted value applied to each tag information.