Wavelet Support Vector Machine and Particle Swarm Optimizer for Prediction of Protein Structural Class

Determination of protein structural class is a quite meaningful topic in protein science. In this paper a wavelet support vector machine (WSVM) coupled with particle swarm optimizer (PSO) is presented for prediction of protein structural class, which is featured by introducing wavelet as a kernel and using PSO to optimize kernel parameters. As a demonstration, the rigorous jackknife cross-validation test was performed on two working datasets that contain 204 and 1673 proteins, respectively. Our success rates were very satisfying, and the optimal mother wavelet was also determined.