Feasibility research on rapid detection of dimethoate in water by near-infrared spectroscopy

Abstract

Near infrared (NIR) spectroscopy combined with chemometrics is used for the rapid detection of dimethoate in water without sample preparation. In order to overcome the non-linear phenomenon least-squared support vector machines (LS-SVM) is utilized. Classification accuracy of the training set and the testing set were 100% when two parameters, γ and σ² (270.191 and 29.60, respectively), were chosen by the Simplex optimizing algorithm. The entire operating procedure, which takes less than 2 minutes per sample to complete, is rapid, convenient and environmentally acceptable. It shows potential for the rapid screening of dimethoate in water samples.