Boosting for real and functional samples: an application to an environmental problem

摘要

In this paper, boosting techniques are given in order to forecast SO_2 levels near a power plant. We use boosting with neural networks to forecast real values of SO_2 concentration. Then, the data are considered as a time series of curves. Assuming a lag one dependence, the predictions are computed using the functional kernel and the linear autoregressive Hilbertian model. Boosting techniques are developed for those functional models. We compare results of functional boosting with different starting points and iterate models. We carry out the estimation, in real and functional cases, with the information given by a historical matrix, which is a subsample that emphasizes relevant SO_2 values.