The boosting technique using correlation coefficient to improve time series forecasting accuracy

Time series forecasting has been considered an important tool to support decisions in different domains. A highly accurate prediction is essential to ensure the quality of these decisions. Time series forecasting is based on historical data and the predictions are usually made using statistical methods. These characteristics make the forecasting problem an interesting application of machine learning techniques, especially for boosting techniques and genetic programming. Boosting techniques currently receive a lot of attention; they combine predictions from different forecasting methods as a procedure to improve the accuracy. This paper explores genetic programming (GP) and boosting technique to obtain an ensemble of regressors and proposes a new formula for the updating of the weights and for the final hypothesis. This new formula is based on the correlation coefficient instead of the loss function used by traditional boosting algorithms, this new algorithm is called boosting using correlation coefficient (BCC). To validate this method, experiments were accomplished using real, financial and artificial series generated by Monte Carlo simulation. The results obtained by using this new methodology were compared with the results obtained from GP, GPBoost and the traditional statistical methodology (ARMA). The results show advantages in the use of the proposed approach.