Abstract
The performance of machine learning such as support vector machines and radial basis function neural networks depends on parameters, for example the width of Gaussian function and a regularization parameter. One of most popular methods for estimating parameters is cross validation (CV) test, however CV test is usually time consuming. In this research, we propose an effective learning method using bagging and boosting in order to reduce the burden on choosing the width of Gaussian function as well as the sensitivity to it. Additionally, we show that comparing with the calculation time by a single machine, the one by the proposed learning method can be improved without losing a performance of generalization ability through several numerical experiments.
