Herbs recognition based on chemical properties using machine learning algorithm

Nur Fadzilah Mohamad Radzi; Azura Che Soh; Asnor Juraiza Ishak; Mohd Khair Hassan.

For decades, the headspace Gas Chromatography Mass Spectrometry (GCMS) technique has been employed to analyse Volatile Organic Compounds (VOCs), extracting chromatographic signals and identifying chemical components. In practical scenarios, identifying major chemical compounds has been a useful approach for herb experts to recognize and differentiate species. However, this process has been manual and lacked an automated herb recognition system that incorporates GCMS technology. To address this gap, a GCMS herb recognition system has been proposed, integrating the GCMS system with a pattern recognition approach. Innovatively, a new feature extraction method using the Weighted Histogram Analysis Method (WHAM) has been introduced. This method employs a reweighting technique that utilizes the peak area and peak height of VOCs to generate a unique pattern for each herb species. A comparison of classification performance between systems with WHAM shows that the Support Vector Machine (SVM) method achieves a higher percentage of accuracy, ranging from 92.32% to 95.67%, compared to without WHAM, which achieves an accuracy ranging from 57.43% to 62.11%. This method has demonstrated promising results in identifying herb species, and the classification method based on machine learning algorithms has proven successful in recognizing and distinguishing herb species.

KEYWORDS: Herbs recognition; Chemical properties; Classification; Machine learning algorithm; Feature extraction.



Download this PDF file

REFERENCES

1. Ahmad, R., Baharum, S., Bunawan, H., Lee, M., Mohd Noor, N., Rohani, E. R., Ilias, N. & Zin, N. M. 2014. Volatile Profiling of Aromatic Traditional Medicinal Plant, Polygonum minus in Different Tissues and Its Biological Activities. Molecules, 19, 19220-19242.
2. Huang, S., Cai, N., Pedro, P.P., Narrandes, S., Wang Y. & Xu, W. 2018. Applications of support vector machine (SVM) learning in cancer genomics. Cancer Genomics & Proteomics, 15(1), 41-51.
3. Ichim, M. C. & Booker, A. 2021. Chemical authentication of botanical ingredients: A review of commercial herbal products. Frontiers in Pharmacology, 12, 1-130.
4. Kumar, A. 2020. Phytochemistry, pharmacological activities and uses of traditional medicinal plant Kaempferia galanga L. – An overview. Journal of Ethnopharmacology, 253, 112667.
5. Kumar, S., Bouzida, D., Swendsen, R. H., Kollman, P. A. & Rosenberg, J. M. 1992. The weighted histogram analysis method for free-energy calculations on biomolecules. Journal of Computational Chemistry, 13(8), 1011-1021.
6. Rana, P., Liaw, S. Y., Lee, M. S. & Sheu, S. C. 2021. Discrimination of four Cinnamomum species with physico-functional properties and chemometric techniques: Application of PCA and MDA models. Foods, 10(11), 2871.
7. Wang, Z., Chen, W., Gu, S., Wang, Y. & Wang, J. 2020. Evaluation of trunk borer infestation duration using MOS E-nose combined with different feature extraction methods and GS-SVM. Computers and Electronics in Agriculture, 170, 105293.
8. Yan, X. & Jia, M. 2018. A novel optimized SVM classification algorithm with multi-domain feature and its application to fault diagnosis of rolling bearing. Neurocomputing, 313, 47-64
9. Zheng, Y. B., Zhang, Z. M., Liang, Y. Z., Zhan, D. J., Huang, J. H., Yun, Y. H. & Xie, H. L. 2013. Application of fast Fourier transform cross-correlation and mass spectrometry data for accurate alignment of chromatograms. Journal of Chromatography, 1286, 175-182.
10. Zouaoui, N., Chenchouni, H., Bouguerra, A., Massouras, T. & Barkat, M. 2019. Characterization of volatile organic compounds from six aromatic and medicinal plant species growing wild in North African drylands. NFS Journal, 18, 19-28.