Research And Experiment On Semantics-based Music Search Method

Music is a kind of art, which formed by organized musical to express the image of the person's thoughts and feelings, to reflect real life. Music is not only an art, it is also a human social phenomenon accompanied by the emergence of human. Today in the highly-developed technological society, music as an essential product of human society, it is booming quickly. Currently, the Internet has stored a great number of music files and it is difficult to conclude precise statistics. Facing the massive music files, how to meet the needs of the user's search is the problem that we need to solve. Currently all the major music search engine to retrieve text information as the main basis for this has been difficult to meet the needs of human intelligent retrieval. With the human semantic technology continues to be researched and explored, Semantic Web technology nearly becomes mature, which marks that based-on natural language search will gradually be on the stage. How can users search for music's natural language? "Music is a vibration with sound waves which exists in the time display, through the human hearing organ caused by a variety of emotional responses and emotional experience of arts [1]." How to make the user's find the music which can retrieve meet their own emotional volatility?Music is an art to express human emotion, and it is also closely related to a social and human phenomenon. Semantic-based music retrieval is mainly to solve the natural language of mankind's search needs. This major work done: (1) Depth analysis of the emotional characteristics of human relations with music, build music ontology. (2) Search on user input when parsing natural language.①Use the word system to query user input word segmentation and POS tagging;②Based retrieval feature extraction keyword collections. (3) Mapping the keywords which obtained by the syntax analysis to the ontology concept, and use the "synonym words forest" for the dictionary for words and sentences semantic similarity calculation. (4) Use inference engine based on description logic for reasoning in music area to meet user's searching requirements. (5) Implemented the prototype of semantics-based music retrieval system. Experimental results show that the system can meet the user's natural language search needs.