A NOVAL AUDIO CLASSIFICATION ALGORITHM BASED ON GA AND SVM WITH COMBINED KERNEL FUNCTION

摘要

Audio classification is an important access to extract audio structure and content, and is a premise for audio content analysis, retrieval and further treatment. Support Vector Machine (SVM) is a valid statistic learning method. In this paper, learning algorithm of SVM is introduced to construct classifier, and construct a new kernel function and use Genetic Algorithm (GA) to optimize the parameters of classifier model. This paper proposes a new audio classification algorithm, GA-CBSVM based on GA and SVM with combined kernel function to classify speech, music and their mixed audio. The experimental results show that GA-CBSVM is excellent for audio classification and the average of classification accuracy is up to 93.08%.