Title :
Bayesian neural networks for electric load forecasting
Author :
Tito, Edison H. ; Zaverucha, Gerson ; Vellasco, Marley ; Pacheco, Marco
Author_Institution :
COPPE, Univ. Fed. do Rio de Janeiro, Brazil
Abstract :
The authors apply Bayesian neural networks to electric load forecasting with real data from some Brazilian power companies. The Bayesian methods used are the Gaussian approximation and the Markov chain Monte Carlo (MCMC) methods. The results obtained with these methods are favourably compared to backpropagation and some standard statistical techniques like Box & Jenkins and Holt-Winters
Keywords :
Bayes methods; Gaussian distribution; Markov processes; Monte Carlo methods; load forecasting; neural nets; power engineering computing; Bayesian methods; Bayesian neural networks; Brazilian power companies; Gaussian approximation; MCMC; Markov chain Monte Carlo; backpropagation; electric load forecasting; real data; standard statistical techniques; Backpropagation; Bayesian methods; Gaussian approximation; Gaussian distribution; Load forecasting; Monte Carlo methods; Neural networks; Predictive models; Probability; Training data;
Conference_Titel :
Neural Information Processing, 1999. Proceedings. ICONIP '99. 6th International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-5871-6
DOI :
10.1109/ICONIP.1999.844023
