ABSTRACT
Many attempts have been carried out to predict dew yield from measurable parameters such as air temperature, relative humidity, wind speed, cloud coverage or radiometric measurements, etc. The models are either based on an energy balance model or a statistical fit of data using artificial intelligence. Artificial neural networks have been widely publicized as a form of artificial intelligence, including various capabilities such as pattern recognition, classification, and predictions. Francl and Panigrahi employed an original approach based on artificial neural networks models using standard meteorological data. Having chosen the architecture of the network, the calculations of weights and bias are optimized by a correct division of the data set. Cooling by radiation deficit between the surface and the atmosphere is only balanced with heating by conduction, convection, and condensation.
