Author :
Bishop, Christopher M. ; Hinton, Geoffrey E. ; Strachan, Iain G D
Author_Institution :
Neural Comput. Res. Group, Aston Univ., Birmingham, UK
Abstract :
The standard GTM (generative topographic mapping) algorithm assumes that the data on which it is trained consists of independent, identically distributed (i.i.d.) vectors. For time series, however, the i.i.d. assumption is a poor approximation. In this paper we show how the GTM algorithm can be extended to model time series by incorporating it as the emission density in a hidden Markov model. Since GTM has discrete hidden states we are able to find a tractable EM algorithm, based on the forward-backward algorithm, to train the model. We illustrate the performance of GTM through time using flight recorder data from a helicopter
Keywords :
neural nets; GTM algorithm; GTM through time; discrete hidden states; emission density; flight recorder data; forward-backward algorithm; generative topographic mapping algorithm; helicopter; hidden Markov model; independent identically distributed vectors; neural network training; performance; time series;
Conference_Titel :
Artificial Neural Networks, Fifth International Conference on (Conf. Publ. No. 440)
Conference_Location :
Cambridge
Print_ISBN :
0-85296-690-3
DOI :
10.1049/cp:19970711
