A GA-BP method of detecting carbamate pesticide mixture based on three-dimensional fluorescence spectroscopy

In this paper, we have proposed a method to detect a mixture of carbamate pesticides using a back propagation network (BP), which is optimized by genetic algorithm (GA) for quantitative analysis. This method aims to combine the advantages of BP and GA to remedy their drawbacks. The training samples were taken as input, some performance indexes such as the predicted values, iteration time, mean squared error, correlation coefficient and recovery rate were compared between BP neural network and the constructed GA-BP model to evaluate the performance of two neural networks. Results show that the optimized GA-BP model can effectively predict the concentrations, the mean squared error and recovery rate are better. In addition, the correlation coefficient has a significant improvement. This study can provide a new way for detection of the pesticides mixture and help to analysis in a reliable way.