caa a22 4500 477034489 CHVBK 20180405111301.0 cr unu---uuuuu 170330e19960601xx s 000 0 eng 10.1007/BF00571683 doi (NATIONALLICENCE)springer-10.1007/BF00571683 Use of modular architectures for time series prediction [Elektronische Daten] [Samy Bengio, Françoise Fessant, Daniel Collobert] Recently, a lot of papers have been published in the field of time series prediction using connectionist models. Nevertheless we think that one of the major problem with is rarely treated in the literature is related to the choice of input parameters (embedding dimension and delay). In this paper, we propose two modular approaches to this problem and apply them to a sunspot-related time series. Experimental results are then compared to a single multi-layer perceptron in order to estimate performances of these models. Kluwer Academic Publishers, 1996 embedding dimension nationallicence mixtures of experts nationallicence modular architectures nationallicence time-series prediction nationallicence Bengio Samy Dept. IRO, Université de Montréal, Succ. Centre-ville, C.P. 6128, H3C 3J7, Montréal, Québec, Canada aut Fessant Françoise LEIBNIZ-IMAG, 46 avenue Felix Viallet, 38031, Grenoble Cédex, France aut Collobert Daniel LAB/RIO/TNT, France Télécom CNET, 22307, Lannion Cédex, France aut Neural Processing Letters Kluwer Academic Publishers 3/2(1996-06-01), 101-106 1370-4621 3:2<101 1996 3 11063 https://doi.org/10.1007/BF00571683 text/html Onlinezugriff via DOI 1 research-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/BF00571683 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Bengio Samy Dept. IRO, Université de Montréal, Succ. Centre-ville, C.P. 6128, H3C 3J7, Montréal, Québec, Canada aut NATIONALLICENCE 700 1- Fessant Françoise LEIBNIZ-IMAG, 46 avenue Felix Viallet, 38031, Grenoble Cédex, France aut NATIONALLICENCE 700 1- Collobert Daniel LAB/RIO/TNT, France Télécom CNET, 22307, Lannion Cédex, France aut NATIONALLICENCE 773 0- Neural Processing Letters Kluwer Academic Publishers 3/2(1996-06-01), 101-106 1370-4621 3:2<101 1996 3 11063 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer