Title :
Nonlinear time series prediction weighted by marginal likelihoods: a hierarchical Bayesian approach
Author :
Matsumoto, T. ; Saito, M. ; Sugi, J.
Author_Institution :
Dept. of Electr., Electron. & Comput. Eng., Waseda Univ., Tokyo, Japan
Abstract :
A nonlinear time series prediction scheme is proposed with a combination of model dynamical systems weighted by marginal likelihoods. The scheme outperforms prediction with a single model prediction with the highest marginal likelihood
Keywords :
Bayes methods; multilayer perceptrons; nonlinear dynamical systems; parameter estimation; time series; hierarchical Bayesian algorithm; marginal likelihood; multilayer perceptron; nonlinear dynamical systems; parameter estimation; time series prediction; Bayesian methods; Distributed computing; Equations; Markov processes; Neural networks; Noise level; Nonlinear dynamical systems; Predictive models; Uncertainty; Yttrium;
Conference_Titel :
Neural Networks, 1999. IJCNN '99. International Joint Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5529-6
DOI :
10.1109/IJCNN.1999.833486
