Assessment of soil pollution hazard by using fuzzy neural network

A methodology for the rapid and efficient assessment of the hazard of soil contamination due to the presence of uncontrolled landfills was carried out in this study. In particular, by means of a fuzzy neural network, the criterion adopted allowed us to compare the results obtained from a cross analysis of some intrinsic characteristics of the single landfills and the territory where they were located. Their identification shows the most relevant environmental problem. Therefore, we have classified each site within a hazard scale enabling us to understand which one requires to be checked more urgently, to do instrumental surveys and, if needed, to do restoration and reclamation. The fuzzy neural network was combined with a sensitivity analysis to identify the best membership function that represents the input data. The proposed fuzzy neural network, supported by the sensitivity analysis, has revealed to be an important tool for supporting decisions, in order to optimise technical and economic resources.