EFFECTIVE AUDIO CLASSIFICATION ALGORITHM USING SWARM-BASED OPTIMIZATION

The effectiveness and usefulness of large audio databases is greatly dependent on the ability to classify and retrieve audio files based on their properties or content. Automatic classification using machine learning is much more practical than manual classification. In this paper, a new audio classification algorithm using Simplified Swarm Optimization (SSO) based on Particle Swarm Optimization (PSO) is presented. The performance of the new algorithm is compared with two existing state-of-the-art classifiers, PSO and Support Vector Machine (SVM), for an audio dataset being classified into five classes of musical instruments. The experimental results show that the proposed SSO-based classifier has improved classification accuracy (91.7%) when compared with PSO (87.2%) and SVM (88.5%). Additionally, the algorithm is shown to have simpler particle update calculations than PSO, and also requires fewer particles for classification training.