Use of Support Vector Machines through Linear-Polynomial (LP) Kernel for Speech Recognition

The kernel functions are playing a very important role in machine learning. In this paper, the speech recognition problem is considered as a machine learning problem. The new kernel function called Linear-Polynomial kernel (LP) used to design the support vector machines for speech recognition for improving the generalization performance of speech recognition. The LP kernel performance is very good as compared to linear kernel and polynomial kernel and has improved the generalization performance ability of the speech recognition system. The One-versus-One approach is used for improving the systems efficiency.