Daily insolation forecasting using a multi-stage neural network

So far a single-stage neural network has been proposed to forecast the inso lation of the next day. The mean error of the forecast insolation by the si ngle-stage neural network is about 30%. In this paper, a multi-stage neural network is developed for further reduction of the mean error. A first-stag e neural network forecasts the average atmospheric pressure of the next day from atmospheric pressure data of the previous day. A second-stage neural network forecasts the insolation level of the next day from the average atm ospheric pressure and weather data of the previous day. A third-stage neura l network forecasts the insolation of the next day from the insolation leve l and weather data of the previous day. Meteorological data at Omaezaki, Ja pan in 1988-1993 are used as input data, and the insolations in 1994 are fo recast. The insolations forecast by the multi-stage and the single-stage ne ural networks are compared with the measured ones. The results show that th e mean error reduces from about 30% (by the single-stage) to about 20% (by the multi-stage). (C) 1999 Elsevier Science Ltd. All rights reserved.