ANALYSIS THE MUSICAL EMOTION BASED ON DATA MINING

Musical emotion is of import for the listener’s cognition. Fast natural activity of digital music data in the Internet during the recent years has led to increase of user demands for search based on different types of meta data. One kind of Meta data that this system focused in this project is the emotion or mood of music. Music emotion acknowledgement is a prevalent research topic today. This instrumentation collected a database including 280 pieces of popular music with four basic emotions of Thar’s two Magnitude model. This system used a two level classifier the process of which could be briefly summarized in three steps: 1) Extracting most suitable features from composition of music in the database to exposit each music song; 2) Applying feature selection coming to decrease correlations between features; 3) Using SVM classifier in two level to train these features. Ultimately this system magnified accuracy rate from 72.14% with simple SVM to 87.27% with our hierarchical classifier.