Resolving Organic Pollutants Based on a Intelligent Computing Method

This paper suggests a novel method named DOSCWTGRNN, which is based on generalized regression neural network (GRNN) with direct orthogonal signal correction (DOSC) and wavelet transform (WT) as a preprocessing tool for the simultaneous spectrophotometric determination of o-nitro-aniline, m-nitro-aniline and p-nitro- aniline. DOSC was applied to remove structured noise that is unrelated to the concentration variables. Wavelet representations of signals provide a local time-frequency description, thus in the wavelet domain, the quality of noise removal can be improved. GRNN was applied for overcoming the convergence problem met in back propagation training and facilitating nonlinear calculation. In this case, by optimization, the number of DOSC components, tolerance factor, wavelet function, decomposition level and the width (sigma) of GRNN for DOSCWTGRNN were selected as 1, 0.001, Coiflet 1, 5 and 0.4 respectively. The relative standard errors of prediction (RSEP) for all components with DOSCWTGRNN, WTGRNN and GRNN were 4.37%, 4.93% and 6.56% respectively. The proposed method has been successfully applied to analyze overlapping spectra and was proven to be better than other techniques.