Meteorological time series forecasting based on MLP modelling using heterogeneous transfer functions

3rd International Conference on Mathematical Modeling in Physical Sciences

物理学;数学

Voyant, C.^1 ; Nivet, M.L.^1 ; Paoli, C.^2 ; Muselli, M.^1 ; Notton, G.^1

University of Corsica-SPE CNRS UMR 6134, Corte-(Corsica)

20250, France^1

University of Galatasaray-Computer Engineering Department, No:36 34357 Ortaköy, Istanbul, Turkey^2

关键词: Hidden layers;    Meteorological data;    Multi layer perceptron;    Seasonal time series;    Time index;    Time series forecasting;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/574/1/012064/pdf
DOI  :  10.1088/1742-6596/574/1/012064

来源: IOP

【 摘 要 】

In this paper, we propose to study four meteorological and seasonal time series coupled with a multi-layer perceptron (MLP) modeling. We chose to combine two transfer functions for the nodes of the hidden layer, and to use a temporal indicator (time index as input) in order to take into account the seasonal aspect of the studied time series. The results of the prediction concern two years of measurements and the learning step, eight independent years. We show that this methodology can improve the accuracy of meteorological data estimation compared to a classical MLP modelling with a homogenous transfer function.