基于最小二乘支持向量机的苯麻滴鼻液含量测定

Determination of Diphenhydramine Hydrochloride and Ephedrine Hydrochloride in Nasal Drops by Least Squares Support Vector Machine

应用最小二乘支持向量机(LS-SVM)技术同时测定苯麻滴鼻液中盐酸苯海拉明和盐酸麻黄碱的含量。对苯麻滴鼻液的紫外光谱数据进行预处理和主成分分析后不经分离,采用LS-SVM学习建模,并同时测定两组分的含量。对均匀设计标样中的盐酸苯海拉明、盐酸麻黄碱拟合相关系数分别为0.9999、0.9993。测试样品误差范围-0.81%～3.68%,测定样品平均回收率分别100.0%、99.7%。与HPLC法所得结果无显著差异。

A least squares support vector machine （LS-SVM） was utilized for simultaneous determination of diphenhydramine hydrochloride and ephedrine hydrochloride in Benrna nasal drops. After UV data were pretreated and the principal component analyzed, LS-SVM was employed to analysis these two components with a high degree of spectral overlap. The coefficient correlations of these two components were 0.9999 and 0.9993 after LS-SVM trained. The error range of prediction was -0.81% - 3.68%, and the average recovery were 100.0% and 99.7%. The results showed that there was no significant difference to HPLC method.