Determination Of Rifampicin And Isoniazide Tablets With Nirs Based On Support Vector Machine

Near Infrared Spectroscopy technology is a fast, efficient, non-polluting testing technology. Near Infrared Spectroscopy technique used in drug analysis is considered an important progress in the analysis of the drug. Support Vector Machine (SVM) has strong nonlinear and anti-interfering capabilities. This paper introduces SVM, a young branch of statistical learning theory, to drug analysis field by the Near Infrared Spectroscopy method.This paper take anti-TB drugs Isoniazid Tablets as the research object. Analyzing the Isoniazid Tablets near-infrared spectroscopy through SVM method, we establish a quantitative analysis model, which provide a tool to rapid test Isoniazid Tablets major components. Comparing with several models, we find the model based on the fast Fourier transform approach to spectrum (FFT-NIR) is superior to other model. As a result of model predicting Isoniazid Tablets major components, the correlation coefficient (R) of isoniazid and rifampin are 0.95037 and 0.95634, the root mean square error of calibration (RMSECV) are 0.01362 and 0.0296.therefore, this method has higher stability and reliability.The results show that the feasibility of near infrared spectrum quantitative analysis model based on SVM method, provide a regression correction method for Near Infrared Spectroscopy analysis drug production procedure and quality control.