Correction model based ANN modeling approach for the estimation of radon concentrations in Ohio

摘要: According to National Cancer Institute, radon is one of the major causes for lung cancer related deaths after smoking in US. To prevent due inhalation there a need determine level concentration each locality, example, zip code and this would help ease identification areas with high thereby allowing necessary preventive measures be taken. However, factors like inapproachability hinder process estimating some places. In such places it common practice estimate concentrations using several interpolation techniques. article, new approach that improves accuracy neural model sensitivity-based correction modeling Ohio proposed. The results are compared commonly used techniques as kriging, radial basis function (RBF), inverse distance weighting (IDW), global polynomial (GPI), local (LPI), recently developed conventional ANN approach. Further, accuracies all above models evaluated based on Willmott's Index ranked performance criteria emphasis extreme-end (peak-end, low-end), mid-range concentrations. demonstrate effectiveness proposed percentage improvement 70–80% prediction accuracy, other © 2012 American Institute Chemical Engineers Environ Prog, 32: 1223–1233, 2013