Multiple Gaussian process models for direct time series forecasting
Tomohiro Hachino, Visakan Kadirkamanathan · IEEJ Transactions on Electrical and Electronic Engineering · 2011
Abstract This paper focuses on the problem of time series forecasting using the Gaussian process models. The Gaussian process model is a nonparametric model and the output of the model has Gaussian distribution with mean and variance. The multiple Gaussian process models as every step ahead predictors are used for time series forecasting in accordance with the direct approach. The separable least‐squares approach that combines the genetic algorithm with the linear least‐squares method is applied to train these Gaussian process models. Simulation results are shown to illustrate the effectiveness of the proposed direct forecasting method. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.