Two Models Of Parameter Selection In V-Support Vector Machine

With the development of modern informational technology, a large number of databases have appeared. Consequently, how to find useful information from sets of data becomes a problem need to be solved imminently. Data mining technology comes into being in this background. Support vector machine is a new technology of data mining, and it has been successfully applied to many fields.Support vector machine(SVM) is a machine learning method that developed on the basis of statistical learning theory on the mid-1990s. It contains the largest margin hy-perplane, Mercer kernal, convex quadratic programming, slack variables and such technologies. Until now it has got the best properties in some challengeable applications.This paper mainly studies the optimal methods for parameter selection in v-support vector machine[v-SVM). The v-support vector machine is a new class of support vector algorithms. v-SVM was proposed by incorporating a change from C in the original SVM algorithm with v which possesses the following intuitive meaning: v is both an upper bound on the fraction of margin errors and a lower bound on the fraction of support vectors. The parameter selection in support vector machine is an important research topic, different parameters result in different generalization. The study of parameter selection in v-SVM is fewer. The parameter selection model for v in v-SVM is presented and the numerical experiment is given in this paper.At first, a model for selecting optimal parameter v in v-SVM is presented in this paper, which bases on mathematical programming with equilibrium constraints. The model is a nonlinear problem which has smoothed objective function and with complementary constrains. Numerical experimental results show that this model is effective in choosing the parameter v. Besides, we discuss the v-SVM when the data sets are unbalanced. An improved 2v-SVM is presented in this paper, this model's merit is the parameter easy to select. Numerical experimental results show that this model is effective when solving unbalanced question.