Title :
Time Series Forecasting with PSO-Optimized Neural Networks
Author :
Alba-Cuellar, Daniel ; Munoz-Zavala, Angel Eduardo ; Ponce-de-Leon-Senti, Eunice Esther ; Diaz-Diaz, Elva ; Hernandez-Aguirre, Arturo
Author_Institution :
Univ. Autonoma de Aguascalientes, Aguascalientes, Mexico
Abstract :
In this paper, we propose a new methodology to forecast values for univariate time series datasets, based on a Feed Forward Neural Network (FFNN) ensemble. Each ensemble element is trained with the Particle Swarm Optimization (PSO) algorithm, this ensemble produces a final sequence of time series forecasts via a bootstrapping procedure. Our proposed methodology is compared against Auto-Regressive Integrated Moving Average (ARIMA) models. This experiment gives us a good idea of how effective soft computing techniques can be in the field of time series modeling. The results obtained show empirically that our proposed methodology is robust and produces useful forecast error bounds that provide a clear picture of a time series´ future movements.
Keywords :
feedforward neural nets; forecasting theory; learning (artificial intelligence); mathematics computing; particle swarm optimisation; time series; FFNN ensemble; PSO-optimized neural networks; bootstrapping procedure; empirical analysis; error bound forecasting; feedforward neural network ensemble; particle swarm optimization algorithm; soft computing techniques; time series forecasting; univariate time series datasets; Biological neural networks; Computational modeling; Forecasting; Predictive models; Time series analysis; Training; ARIMA models; Artificial Intelligence; Artificial Neural Networks; Bootstrapping; Committee Machines; Evolutionary Algorithms; Particle Swarm Optimization; Soft Computing; Time Series Forecasts;
Conference_Titel :
Artificial Intelligence (MICAI), 2014 13th Mexican International Conference on
Print_ISBN :
978-1-4673-7010-3
DOI :
10.1109/MICAI.2014.22