食品安全隐患越来越受到重视, 而食品添加剂的过量使用更是个重要的因素。应用FS920荧光光谱仪, 研究了防腐剂山梨酸钾的荧光特性, 得到山梨酸钾荧光特征峰于λex/λem=375/490nm采用基于最小二乘支持向量机对橙汁溶液中防腐剂山梨酸钾进行检测, 通过改进的遗传算法寻优最小二乘支持向量机参数。经过样本训练得到橙汁溶液山梨酸钾的回归模型, 对未知浓度的溶液进行预测, 将新算法与基本遗传算法寻优的模型和BP神经网络对比。结果表明, 自适应遗传-最小二乘支持向量机建立的预测模型在平均相对误差3.54%和平均回收率96.46%都是最优的,是一种准确有效的橙汁中山梨酸钾浓度检测方案。

More and more attention has been paid to food safety, and the excessive use of food additives is an important issue. the fluorescence characteristic of potassium sorbate is studied by using FS920 fluorescence spectrometer, and the result shows that the fluorescence characteristic peak of potassium sorbate is λex/λem=375/490nm. Least squares support vector machine to detect potassium sorbate in orange juice solution is used. Additionally, the adaptive genetic algorithm is used to optimize the parameters of least squares support vector machine. The regression model of potassium sorbate in orange juice solution is obtained through the sample training, and the solution of the unknown concentration is predicted, and this algorithm is compared with the BP neural network for the optimization of the basic genetic algorithm. The results show that the prediction model established by AGA-LSSVM is the best one with the average relative error of 3.94% and the average recovery rate of 96.18%, Which providing an accurate and effective solution for the concentration detection of potassium sorbate in orange juice.