Makine Öğrenmesi Yöntemleri ile Kenevir Türlerinin THC Seviyelerinin Tahmin Edilmesi

Yıl 2024, Cilt: 3 Sayı: 2, 125 - 136

Talip Çay

Öz

Bu çalışmada, hibrit, sativa ve indica kenevir türlerinin terpen profillerine dayanarak THC (tetrahidrokannabinol) seviyelerinin tahmin edilmesi amacıyla makine öğrenmesi algoritmaları kullanılmıştır. Kenevir bitkisi, tıbbi ve endüstri alanında kullanımlarıyla popülarite kazanmıştır; bu nedenle, bu türlerin özelliklerinin anlaşılması kritik öneme sahiptir. Çalışmada, kenevir türlerinin terpen bileşenleri, THC seviyelerini etkileyen önemli faktörlerden biri olarak belirlenmiştir. Veri seti, farklı kenevir türlerine ait terpen bileşenlerini ve bu türlerin THC seviyelerini içermektedir. Çalışmada regresyon analizi, k-NN (k En Yakın Komşuluk), SVM (Destek Vektör Makineleri) ve YSA (Yapay Sinir Ağları) gibi klasik ve derin öğrenme algoritmaları kullanılarak modeller geliştirilmiştir. Her bir algoritma performans metrikleri değerlendirilmiştir. % 94’e ulaşan doğruluk oranları ile elde edilen sonuçlar, terpen bileşenlerinin THC seviyelerinin tahmininde önemli bir rol oynadığını göstermiştir. Bu çalışma, kenevir türlerinin analiz edilmesinde makine öğrenmesi yöntemlerinin uygulanabilirliğini ortaya koymuş ve terpen bileşenleri ile THC seviyesi arasındaki ilişkileri daha iyi anlamak için bir temel oluşturmuştur.

Anahtar Kelimeler

Makine öğrenmesi, YSA, kNN, doğrusal bağlanım, SVM, THC, terpen, hibrit, sativa, indica, kenevir

Etik Beyan

Çalışmamızda kullanılan veriler kamuya açıktır ve Veri Kullanılabilirliği Beyanımızda ayrıntılı olarak açıklandığı gibi veri gizliliğine tabi değildir. Bu nedenle, bu verilerin kullanımı için etik onayı gerekliliği yoktur. (Cannabis Strains Dataset. Kaggle. https://www.kaggle.com/datasets/corykjar/leafly-cannabis-strains-dataset [Access date: October 10, 2024])

Prediction of THC Levels in Cannabis Species Using Machine Learning Methods

Öz

In this study, machine learning algorithms were utilized to predict THC (tetrahydrocannabinol) levels based on the terpene profiles of hybrid, sativa, and indica cannabis species. The cannabis plant has gained popularity due to its uses in medical and industrial fields; therefore, understanding the characteristics of these species is of critical importance. In the study, the terpene components of cannabis species were identified as one of the significant factors affecting THC levels. The dataset includes terpene components of different cannabis species and the THC levels of these species. Models were developed using classical and deep learning algorithms, such as regression analysis, k-NN (k-Nearest Neighbors), SVM (Support Vector Machines), and ANN (Artificial Neural Networks). The performance metrics of each algorithm were evaluated. The results, with accuracy rates reaching 94%, indicate that terpene components play a significant role in predicting THC levels. This study demonstrates the applicability of machine learning methods in analyzing cannabis species and establishes a foundation for better understanding the relationships between terpene components and THC levels

Anahtar Kelimeler

Machine Learning, ANN, kNN, linear regression, SVM, THC, Terpene, Hybrid, Sativa, Indica, cannabis

Etik Beyan

The data used in our study are publicly available and not subject to data privacy, as detailed in our Data Availability Statement. Therefore, there is no requirement for ethical approval for the use of these data. (Cannabis Strains Dataset. Kaggle. https://www.kaggle.com/datasets/corykjar/leafly-cannabis-strains-dataset [Access date: October 10, 2024])

Kaynakça

• Abeysekera, S.K. et al. (2023) ‘Sparse reproducible machine learning for near infrared hyperspectral imaging: Estimating the tetrahydrocannabinolic acid concentration in Cannabis sativa L.’, Industrial Crops and Products, 192, p. 116137. Available at: https://doi.org/10.1016/J.INDCROP.2022.116137.
• Andre, C.M., Hausman, J.F. and Guerriero, G. (2016) ‘Cannabis sativa: The plant of the thousand and one molecules’, Frontiers in Plant Science, 7(FEB2016), p. 174167. Available at: https://doi.org/10.3389/FPLS.2016.00019/BIBTEX.
• Awad, M. and Khanna, R. (2015) ‘Support Vector Regression’, Efficient Learning Machines, pp. 67–80. Available at: https://doi.org/10.1007/978-1-4302-5990-9_4.
• Birenboim, M. et al. (2022) ‘Use of near-infrared spectroscopy for the classification of medicinal cannabis cultivars and the prediction of their cannabinoid and terpene contents’, Phytochemistry, 204, p. 113445. Available at: https://doi.org/10.1016/J.PHYTOCHEM.2022.113445.
• Campbell, L.G., Peach, K. and Wizenberg, S.B. (2021) ‘Dioecious hemp (Cannabis sativa L.) plants do not express significant sexually dimorphic morphology in the seedling stage’, Scientific Reports 2021 11:1, 11(1), pp. 1–8. Available at: https://doi.org/10.1038/s41598-021-96311-w.
• Casano, S. et al. (2011) ‘Variations in terpene profiles of different strains of Cannabis sativa L.’, Acta Horticulturae, 925, pp. 115–122. Available at: https://doi.org/10.17660/ACTAHORTIC.2011.925.15.
• Costa, B. (2007) ‘On the Pharmacological Properties of Δ9-Tetrahydrocannabinol (THC)’, Chemistry & Biodiversity, 4(8), pp. 1664–1677. Available at: https://doi.org/10.1002/CBDV.200790146.
• Crawford, S. et al. (2021) ‘Characteristics of the Diploid, Triploid, and Tetraploid Versions of a Cannabigerol-Dominant F1 Hybrid Industrial Hemp Cultivar, Cannabis sativa “Stem Cell CBG”’, Genes 2021, Vol. 12, Page 923, 12(6), p. 923. Available at: https://doi.org/10.3390/GENES12060923.
• Geskovski, N. et al. (2021) ‘Mid-infrared spectroscopy as process analytical technology tool for estimation of THC and CBD content in Cannabis flowers and extracts’, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 251, p. 119422. Available at: https://doi.org/10.1016/J.SAA.2020.119422.
• Grotenhermen, F. and Russo, E. (2013) ‘Cannabis and Cannabinoids: Pharmacology, Toxicology, and Therapeutic Potential’, Cannabis and Cannabinoids: Pharmacology, Toxicology, and Therapeutic Potential, pp. 1–439. Available at: https://doi.org/10.4324/9780203479506/CANNABIS-CANNABINOIDS-ETHAN-RUSSO-ETHAN-RUSSO.
• Hanuš, L.O. and Hod, Y. (2020) ‘Terpenes/Terpenoids in Cannabis: Are They Important?’, Medical Cannabis and Cannabinoids, 3(1), pp. 25–60. Available at: https://doi.org/10.1159/000509733.
• Hastie, T., Tibshirani, R. and Friedman, J. (2009) ‘The Elements of Statistical Learning’. Available at: https://doi.org/10.1007/978-0-387-84858-7.
• Hitziger, S. and Modeling, S.H. (2015) ‘Modeling the variability of electrical activity in the brain’. Available at: https://theses.hal.science/tel-01175851 (Accessed: 29 October 2024).
• James, G. et al. (2023) ‘An Introduction to Statistical Learning’. Available at: https://doi.org/10.1007/978-3-031-38747-0.
• Jennings, P.R. (1966) The Evolution of Plant Type in Oryza sativa on JSTOR, Economic Botany. Available at: https://www.jstor.org/stable/4252799 (Accessed: 29 October 2024).
• LaVigne, J.E. et al. (2021) ‘Cannabis sativa terpenes are cannabimimetic and selectively enhance cannabinoid activity’, Scientific Reports 2021 11:1, 11(1), pp. 1–15. Available at: https://doi.org/10.1038/s41598-021-87740-8.
• Livingston, S.J. et al. (2020) ‘Cannabis glandular trichomes alter morphology and metabolite content during flower maturation’, The Plant Journal, 101(1), pp. 37–56. Available at: https://doi.org/10.1111/TPJ.14516.
• Lu, S. et al. (2024) ‘Advances in Machine Learning Processing of Big Data from Disease Diagnosis Sensors’, ACS Sensors, 9(3), pp. 1134–1148. Available at: https://doi.org/10.1021/ACSSENSORS.3C02670/ASSET/IMAGES/MEDIUM/SE3C02670_0010.GIF.
• Lu, Y. et al. (2022) ‘Hyperspectral Imaging With Machine Learning to Differentiate Cultivars, Growth Stages, Flowers, and Leaves of Industrial Hemp (Cannabis sativa L.)’, Frontiers in Plant Science, 12, p. 810113. Available at: https://doi.org/10.3389/FPLS.2021.810113/BIBTEX.
• Pisanti, S. and Bifulco, M. (2019) ‘Medical Cannabis: A plurimillennial history of an evergreen’, Journal of Cellular Physiology, 234(6), pp. 8342–8351. Available at: https://doi.org/10.1002/JCP.27725.
• Radwan, M.M. et al. (2021) ‘Cannabinoids, Phenolics, Terpenes and Alkaloids of Cannabis’, Molecules 2021, Vol. 26, Page 2774, 26(9), p. 2774. Available at: https://doi.org/10.3390/MOLECULES26092774.
• Sommano, S.R. et al. (2020) ‘The Cannabis Terpenes’, Molecules 2020, Vol. 25, Page 5792, 25(24), p. 5792. Available at: https://doi.org/10.3390/MOLECULES25245792.
• Turhan, N. and Yurttakal, A.H. (2022) ‘Regression and Discrimination Performance of Terpenoid Expression via Cannabis Sativa Cannabinoids’, HORA 2022 - 4th International Congress on Human-Computer Interaction, Optimization and Robotic Applications, Proceedings [Preprint]. Available at: https://doi.org/10.1109/HORA55278.2022.9800044.
• Vergara, L.A., Hortúa, H.J. and Orozco, G.A. (2022) ‘Property Estimation Method for Cannabinoids and Terpenes Using Machine Learning’, Computer Aided Chemical Engineering, 51, pp. 103–108. Available at: https://doi.org/10.1016/B978-0-323-95879-0.50018-7.
• Vilela, L.F.S. et al. (2019) ‘Forecasting financial series using clustering methods and support vector regression’, Artificial Intelligence Review, 52(2), pp. 743–773. Available at: https://doi.org/10.1007/S10462-018-9663-X/FIGURES/11.
• Yildirim, S. and Koca Çalişkan (2020) ‘Hemp And Its Use In Health (Kenevir Ve Sağlık Alanında Kullanımı)’, avesis.gazi.edu.tr U Koca Çalışkan J. Fac. Pharm. Ankara/Ankara Ecz. Fak. Derg., 2020•avesis.gazi.edu.tr, 44(1), pp. 112–136. Available at: https://doi.org/10.33483/jfpau.559665.
• Zandkarimi, F. et al. (2023) ‘Comparison of the Cannabinoid and Terpene Profiles in Commercial Cannabis from Natural and Artificial Cultivation’, Molecules, 28(2), p. 833. Available at: https://doi.org/10.3390/MOLECULES28020833/S1.