The establishment and evaluation of near infrared universal model to determinate the effective ingredient content in pesticide rapidly

A near infrared universal quantitative analysis model was established to determinate the effective ingredient content in pesticide EC (hikernalisation) by the PLS (partial least squares) algorithm, the model predictive ability was evaluated by the external inspection method. The model was established among samples containing the same active ingredient from five different companies, and the model determination coefficient R2 and RMSECV (root mean square error of cross validation) were 0.9997 and 0.0223, respectively, the relative error between predicted value and chemical value of the testing set samples was between −2.71% and 3.36%, which indicated that the method to determinate the effective ingredient content in pesticide EC by the established universal model can meet the need of pesticide market monitoring.