Title :
Time-line hidden Markov experts for time series prediction
Author :
Wang, Xin ; Whigham, Peter ; Deng, Da ; Purvis, Martin
Author_Institution :
Dept. of Inf. Sci., Otago Univ., Dunedin, New Zealand
Abstract :
A modularised connectionist model, based on the mixture of experts (ME) algorithm for time series prediction, is introduced. A group of connectionist modules learn to be local experts over some commonly appeared states in a time series. The dynamics for combining the experts is a hidden Markov process, in which the states of a time series are regarded as states of a HMM and each of them associates to an expert. However, the state transition property is time-variant and conditional on the dynamic situation of the time series.
Keywords :
divide and conquer methods; hidden Markov models; learning (artificial intelligence); neural nets; prediction theory; time series; mixture of experts algorithm; modularised connectionist model; state transition property; time series prediction; time-line hidden Markov experts; Data mining; Hidden Markov models; Information science; Markov processes; Noise measurement; Predictive models; State-space methods; Time measurement; Time series analysis; White noise;
Conference_Titel :
Neural Networks and Signal Processing, 2003. Proceedings of the 2003 International Conference on
Conference_Location :
Nanjing
Print_ISBN :
0-7803-7702-8
DOI :
10.1109/ICNNSP.2003.1279393
