QUANTITATIVE ANALYSIS OF Al IN FLOUR PRODUCTS BY LASER-INDUCED BREAKDOWN SPECTROSCOPY COMBINED WITH PARTIAL LEAST SQUARES

Based on partial least squares (PLS) analysis, the effects of different smoothing points and different preprocessing methods on the accuracy and precision of the PLS model were compared and analyzed The results show that the PLS quantitative model has the best effect when using 11-point smoothing combined with standard normal variables (SNVs) as the preprocessing method The coefficients of determination (R(c)2, R-p(2)) of the model training set and prediction set are 0.9900 and 0.9996, respectively. The root means square errors RMSECV and RMSEP are 11.7 and 5.23, respectively, and the average relative error of prediction is only 2.96%, which indicates a high prediction accuracy rate and accuracy. This work demonstrates that LIBS technology has broad application prospects for the quantitative detection of specific ingredients in flour products and provides a basis for the real-time monitoring and evaluation of food safety.