Classification of pluviometric networks located in the region of Bogotá, Colombia using artificial neural networks

Abstract

This work presents a methodology for the classification of pluviometric networks using artificial neural networks. For this, the network of stations registered in the Corporación Autónoma Regional de Cundinamarca, Colombia, was analyzed. The network studied consists of 182 stations for the measurement of precipitation and it has a historical series that goes, in some cases, from 1931 to the present. For the classification, three scenarios called types were proposed, in which the number of neurons in the output layer was varied. It was significant that when comparing the results of the different types, the permanence of certain features in the classification was found, indicating the validity of the classification.