Artificial Neural Network as a Prediction Tool in Agricultural Variables

Artificial Neural Network (ANN) technology is a form of artificial intelligence that learns by processing representative data patterns through its internal architecture. ANN technology often offers a superior alternative to traditional physical-based models, and excel at uncovering patterns or relationships in data. It is also a powerful non-linear estimator which is recommended when the functional form between input-output is unknown or it is not well understood but it is believed could be nonlinear. This paper show two applications of ANN for forecasting solar radiation and available dam storage. In the first case ANN provided a good accurate forecasts for all period with an average square error of 0.05% in the prediction. For the second case, ANN provide relatively accurate estimates of water availability one month into the future in the Purisima dam located in the state of Guanajuato. The results suggest that additional input variables such as runoff, cropping patterns and groundwater extractions may be necessary to increase ANN forecasting accuracy. This feasibility study demonstrates that ANN technology has the potential to serve as a highly accurate forecasting tool. Moreover, ANN technology can continuously be updated, as new data become available, increasing its forecasting ability.