Learning and Data Driver Methods for Short Term Meteo Forecast

In this paper an “intelligent systems” based on neural networks and a statistical non parametric method evaluates the future evolution of meteorological variables and the occurrence of particular phenomena such as rain and road ice. The meteorological variables forecast system is based on a recurrent feed forward multi layer perceptron which make the forecasts using data coming from synoptic electronic sensors. Road ice is forecasted using an analytical model.