Near-infrared diffuse reflectance spectroscopy (NIRDRS) was applied to classification and quantification of azithromycin tablets with the aid of chemometric multivariate analysis. Repeatability was investigated by repeated measurements, and the effect of morphology was examined by preparing the tablets in four forms, i.e. tablet product, tablet without coating, powder of tablet without coating, and powder of tablet. Furthermore, baseline elimination by continuous wavelet transform (CWT) and wavenumber selection was discussed for improving the repeatability and accuracy of the method. The results show that the spectra of the samples in the four forms can be measured with an acceptable repeatability, and classification of manufacture sites and quantitative analysis of the active pharmaceutical ingredient (API) can be achieved by principal component analysis (PCA) and partial least squares (PLS) regression, respectively. More importantly, baseline elimination and wavenumber selection can significantly simplify the calculation and improve the results. (C) 2012 Elsevier B.V. All rights reserved.

• Nankai University

• API

1. API
2. Continuous wavelet transform
3. Near-infrared diffuse reflectance spectroscopy
4. Partial least squares
5. Principal component analysis
6. Research Article
7. Wavenumber selection