Comparison of partial least squares and artificial neural network chemometric techniques in determination of sulfamethoxazole and trimethoprim in pharmaceutical suspension by ATR–FTIR spectrometry

Article type: Research Article

Authors: Khanmohammadi, M.^; | Dallali, N. | Bagheri Garmarudi, A.^; | Zarnegar, M. | Ghasemi, K.

Affiliations: Department of Chemistry, Faculty of Science, Imam Khomeini International University, Qazvin, Iran | Department of Chemistry, Faculty of Science, Zanjan University, Zanjan, Iran | Department of Chemistry and Polymer Laboratory, Engineering Research Institute, Tehran, Iran

Note: [] Corresponding author: M. Khanmohammadi, P.O. Box 34149-1-6818, Qazvin, Iran. Tel./Fax: +98 281 3780040; E-mail: [email protected].

Abstract: Partial Least Square (PLS) and Artificial Neural Network (ANN) techniques were compared during development of an analytical method for quantitative determination of sulfamethoxazole (SMX) and trimethoprim (TMP) in Co-Trimoxazole® suspension. The procedure was based on Attenuated Total Reflectance Fourier Transform Infrared (ATR–FTIR) spectrometry. The 800–2500 cm−1 spectral region was selected for quantitative analysis. R2 and relative error of prediction (REP) in PLS technique were (0.989, 2.128) and (0.986, 1.381) for SMX and TMP, respectively. These statistical parameters were improved using the ANN models considering the complexity of the sample and the speediness and simplicity of the method. R2 and RMSEC in modified method were (0.997, 1.064) and (0.997, 0.634) for SMX and TMP, respectively.

Keywords: Artificial neural network, quantitative analysis, sulfamethoxazole, trimethoprim, ATR–FTIR spectrometry, partial least squares

DOI: 10.3233/SPE-2011-0520

Journal: Spectroscopy, vol. 26, no. 2, pp. 105-114, 2011