Data Mining Application In The Classification Of Music

In order to meet the needs of different groups of people who love music. The music software companies started to focus on data mining in the music classification and soon attracted merchants and attention,For it has great commercial value and broad application prospects.The main work and innovation are as follows:Firstly, a comprehensive introduction to the music classification was mentioned. By contrast at home and abroad, we summarized the latest research achievements in the music classification. Then it introduced the BP neural network, RBF neural network, the LVQ neural network, Hoppfeild neural networks.To explain the different principles followed in the algorithm derivation, it described the difference of its practical application as well as the advantages and disadvantages between them.This paper use the neural network classifier to classify the data provided in this article and explain the reasons for their different classification. In order to solve the problems of selecting parameters of support vector machine (SVM),it introduces genetic algorithm and particle swarm algorithm. It verified that after the optimization the accuracy of classification has been improve obviously.Finally, this paper introduces the traditional methods such as linear discriminant analysis and logistic regression analysis.This is the first use of the particle swarm algorithm to solve the support vector machine parameter selection problem in music classification. Compared with the genetic algorithm they are all set random solution from the beginning and then to find the optimal solution from different directions, but the particle swarm algorithm does not have cross, mutation and other operations, the rules are simple, easy to implement. Comparative analysis shows that optimization has a good classification results in the instance of this music.