紫外-可见漫反射光谱-支持向量回归法快速测定西咪替丁片剂含量

采用紫外-可见漫反射光谱-支持向量回归法（UV-Vis DRS-SVR）建立了快速、无损测定西咪替丁片剂的定量方法。人工配制58个西咪替丁片剂粉末样，分为校正集和预测集。经过光谱预处理、异常值剔除和参数优化，最终采用原始光谱，在220.17 nm-980.80 nm，参数g=0.02, C=20, p=0.03的条件下，建立SVR模型，校正集和预测集的决定系数（R2）分别为0.9782, 0.9871，对5个盲样和15个批次市售西咪替丁片剂的预测均方根差（RMSEP）分别为0.0360, 0.0448，比PLS模型的预测精度更高，效果更好。研究表明，UV-Vis DRS-SVR用于快速、非破坏性测定药物制剂中的活性成分是可行的，有望用于制剂生产过程中的质量控制。

Determination of Cimetidine Tablets by Ultraviolet-Visible Diffuse Reflection Spectrum and Support Vector Regression

A method of rapid and nondestructive determination of cimetidine tablets was established by ultraviolet-visible diffuse reflection spectrum coupled with support vector regression (UV-Vis DRS-SVR). By spectral pretreatment, outlier detection and parameter optimization, SVR model was established in range of 220.17-980.80nm with g = 0.02, C = 20, p = 0.03. The determination coefficients (R2) of calibration set and prediction set were 0.9782 and 0.9871, respectively. Compared with PLS model, the SVR model had better prediction performance for 5 blind samples and 15 commercial cimetidine tablet samples with a RMSEP of 0.0360 and 0.0448, respectively. The results showed that UV-Vis DRS-SVR is feasible to apply for quantitative analysis of active compound in pharmaceutical formulations, being expected to be used for the quality control in the process of production of pharmaceutical formulations.