The Application of Least Squares Support Vector Machine for Classification

摘要

In the training of stand support vector machine, the issue is reformulated and represented in such a way so as to obtain a quadratic programming problem. In order to solve the problems expediently, the Least squares support vector machines (LS-SVMs) was introduced. In which, one uses the equality constraints instead of inequality in the conventional SVMs, in this way the optimal solution can be obtained by solving a set of linear equations instead of solving a quadratic programming problem. In the paper, we imply the LS-SVMs into a real dataset, the results show the good performance of the LS-SVMs.