Artificial neural networks vs regression techniques in the forecasting of contaminants in the Santiago River, based on the sample of a pollutant, through Data Fusion

The present article shows a comparison of the use of both Neuronal Networks and Regression methods in the pollutant forecast in the Santiago River, analyzing data obtained from previous samples by data fusion and considering a base pollutant to predict the relationship with more pollutants present in the river, avoiding the detailed and particular tests of the other Contaminants. This paper determines how to choose the best algorithm of the both, based on the error forecast.