Neuronal network model to predict pollution by urban dust from major passageways in Bogotá, Colombia

Abstract

The use of artificial neuronal network (ANN) involves a limited number of variable to predict the behavior of some phenomenal with very good results. This work is used a ANN model to identify site with high concentration of heavy metal from magnetic parameters. The study was done on urban dust samples coming from of vial net of Bogota City, Colombia. The results indicated an extensive distribution of magnetic material and heavy metal (Cr, Cu, Ni, Pb, V y Zn) in urban dust. It was detected that there are several sites in vial net showed high concentration of heavy metal with values of pollution load index mayor to 3. Several models of artificial neuronal network were tested. We found that architecture of 3, 2 neuronal allow to estimate shape precise the site contaminate through of magnetic parameters (mean quadratic error was 3.14 and correlation coefficient between the real values and estimated values was to 0.60)