Time series regression and prediction based on boosting regression

Abstract

In this paper we propose a boosting regression model for time series using BP network and SVR as basic learning methods. We first make brief introduction on BP network and SVR, then give the specific boosting regression algorithm with theoretical analysis. In the experiment, we use a time series data of wind-speed from a coal mine as a training set to verify the efficiency of our proposed method. The experiment results show that boosting regression gain better performance on test training and generaliz ation.