中文摘要：

为实现鸭蛋蛋清中金霉素残留含量的快速测定与检测模型精度的提高,应用导数同步荧光法结合支持向量回归（SVR）建立鸭蛋蛋清中的金霉素含量的预测模型。首先在280～400 nm光谱范围内对三维同步荧光指纹图谱与三维同步荧光光谱进行分析,确定检测鸭蛋蛋清中的金霉素含量的最佳波长差△λ为70 nm;然后对原始的同步荧光光谱进行一阶导数预处理,并用联合区间偏最小二乘法（siPLS）结合遗传算法（GA）优选出11个特征波长;最后应用支持向量回归（SVR）建立了鸭蛋蛋清中的金霉素含量的预测模型,其模型预测集的决定系数（R2）和预测均方根误差（RMSEP）分别为0.973 mg/L和1.501 1 mg/L。试验结果表明,本研究方法能够满足鸭蛋蛋清中金霉素残留含量快速测定要求。

英文摘要：

To achieve the rapid prediction of aureomycin contents in duck egg white and improve the ac- curacy of prediction model, the prediction model of aureomycin content in duck egg white was developed with derivative synchronous fluorescence method and support vector regression （SVR）. Firstly, three-dimensional synchronous fluorescence fingerprints and three-dimensional synchronous fluorescence speetroseopies within 280 -400 nm were analysed, and 70 nm was selected as the optimum wavelength difference. Secondly, 11 characteristic wavelengths were selected combined with siPLS and GA. Lastly, 11 characteristic wavelengths were used to build prediction model applied by SVR. The determination coefficient （R2） and the root mean squared error of prediction （RMSEP） for prediction model were 0.973 and 1. 501 1 mg/L, respectively. The experimental results showed that the research method in the paper could satisfy the request of determination of aureomycin content in duck egg white.