Abstract:

On the basis of support vector machine algorithm and the result of principal component regression of the linear prediction and radial basis function neural network of the nonlinear prediction, a new forecasting model was proposed by which one can effectively improve the prediction accuracy and
 solve the problem of single prediction. Application of the new prediction model to the prediction of finance data showed that compared with the traditional principal component regression and radial basis neural network method, the new model has better effect and practical significance in prediction.

Key words: principal component regression, radial basis neural network, support vector machine, prediction