An Iterative Learning Algorithm Based on Least Squares Support Vector Regression Machines

Aiming at the problem of the large training data set leading to amounts of calculation a sparse approximation algorithm of least squares support vector machine are proposed. Firstly using the thought of matrix block, convert the optimization problem of a Least Squares support vector machine into low order symmetric positive definite linear systems. Furthermore, use conjugate gradient algorithm which has the number of smaller conditions ,reduce the number of iterations about least squares support vector machine learning process At the same time, we also proved theoretically that the new algorithm is of faster convergence rate. Dramatically improve the speed of the algorithm for learning and prediction. Finally, experimental results show that the algorithm has a very good performance in terms of the accuracy of forecast and speed of training.