TY - GEN
T1 - Classification using localized mixtures of experts
AU - Moerland, Perry
PY - 1999
Y1 - 1999
N2 - A mixture of experts consists of a gating network that learns to partition the input space and of experts networks attributed to these different regions. This paper focuses on the choice of the gating network. First, a localized gating network based on a mixture of linear latent variable models is proposed that extends a gating network introduced by Xu et al., based on Gaussian mixture models. It is shown that this localized mixture of experts model, can be trained with the Expectation Maximization algorithm. The localized model is compared on a set of classification problems, with mixtures of experts having single or multilayer perceptrons as gating network. It is found that the standard mixture of experts with feed-forward networks as gate often outperforms the other models.
AB - A mixture of experts consists of a gating network that learns to partition the input space and of experts networks attributed to these different regions. This paper focuses on the choice of the gating network. First, a localized gating network based on a mixture of linear latent variable models is proposed that extends a gating network introduced by Xu et al., based on Gaussian mixture models. It is shown that this localized mixture of experts model, can be trained with the Expectation Maximization algorithm. The localized model is compared on a set of classification problems, with mixtures of experts having single or multilayer perceptrons as gating network. It is found that the standard mixture of experts with feed-forward networks as gate often outperforms the other models.
UR - http://www.scopus.com/inward/record.url?scp=0033356826&partnerID=8YFLogxK
U2 - https://doi.org/10.1049/cp:19991216
DO - https://doi.org/10.1049/cp:19991216
M3 - Conference contribution
SN - 0852967217
SN - 9780852967218
T3 - IEE Conference Publication
SP - 838
EP - 843
BT - IEE Conference Publication
PB - IEE
T2 - Proceedings of the 1999 the 9th International Conference on 'Artificial Neural Networks (ICANN99)'
Y2 - 7 September 1999 through 10 September 1999
ER -