Prediction of active substance content in pharmaceutical tablets using wavelet compressed NIR and Raman spectroscopic data

C. S. Soh, P. Raveendran

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

1 Citation (Scopus)


Wavelet transformation is performed on NIR transmittance and Raman spectroscopic data followed by prediction of active substance content of pharmaceutical tablet from the spectral data. Partial least squares regression (PLSR) is used to build the prediction models. Comparison is made between prediction models with and without wavelet compression. Results show that wavelet-transformed NIR spectral data extracted from certain wavelet scales could improve the prediction accuracy of the PLSR model. In addition, wavelet transformation of NIR spectral data reduces the number of variables used in calibration and prediction. However, there is no improvement in prediction accuracy for wavelet-transformed Raman spectral data.

Original languageEnglish
Title of host publicationICBPE 2006 - Proceedings of the 2006 International Conference on Biomedical and Pharmaceutical Engineering
Number of pages4
Publication statusPublished - 2006
Externally publishedYes
EventInternational Conference on Biomedical and Pharmaceutical Engineering 2006 - Singapore, Singapore
Duration: 11 Dec 200614 Dec 2006
Conference number: 1st (Proceedings)


ConferenceInternational Conference on Biomedical and Pharmaceutical Engineering 2006
Abbreviated titleICBPE 2006
Internet address


  • Near infrared
  • Pharmaceutical tablet
  • Raman
  • Wavelet

Cite this