Selection of the Parameter in Gaussian Kernels in Support Vector Machine

Support vector machine has become a leading method in classifications and is one of the major topics in supervised learning. Its simple idea and fast implementation have made this method widely used in many areas, such as economics, natural science and chemical engineering. Gaussian kernel is the most common kernel in the support vector machine method, however, the selection of the parameter sigma has not become clear yet. In this paper, we study the selection of sigma based on separation and cohesion. The data is about the major evaluations in Chuzhou Vocational and Technical College. Our second goal in this paper is to determine which majors are performed well and which are not based on support vector machine method.