【摘要】 为了提高多组分混合农药残留荧光检测的含量预测精度，该文提出一种基于连续投影算法(Successive projections algorithm, SPA)和最小二乘支持向量机(Least squares support vector machines, LSSVM)的建模方法。该方法首先应用连续投影算法分别优选出灭蝇胺、异丙甲草胺、克菌丹及噻虫嗪4种农药对应的特征波长，获得25个特征波长点作为预测模型的输入；然后应用最小二乘支持向量机方法对该四组分混合农药残留进行含量预测建模，发现该方法模型的预测精度高于传统偏最小二乘回归模型，验证了方法的有效性。试验结果表明，该方法可用于多组分混合农药残留的荧光检测，且检测精度良好。更多还原

【Abstract】 In order to improve the accuracy of content prediction for multi-component mixed pesticide residues by fluorescence detection, a novel method based on successive projections algorithm(SPA)and least squares support vector machines(LSSVM)is proposed. Firstly, the feature wavelengths of cyromazine, metolachlor, captan and thiamethoxam are selected by successive projection algorithm, and 25 characteristic wavelength points are obtained as the input of the prediction model. Then, the least squares support vector machines method is applied to predict the content of the four pesticide residues in the mixed solution. It is found that the prediction accuracy of the method is higher than that of the traditional partial least squares regression model, which verifies the effectiveness of the method. The experimental results show that the SPA-LSSVM method can be applied for the fluorescence detection of multi-component pesticide residues, and has high prediction accuracy.更多还原