Regression and Discrimination Performance of Terpenoid Expression via Cannabis Sativa Cannabinoids

Cannabis is a herbaceous plant with high economic returns, known for its euphoric and therapeutic properties. It can be used in many industrial areas such as health, medicine, textile, furniture. Organic compounds from essential oils called terpenes that give flavor and aroma to hemp as well as to all plants. Terpenes and cannabinoids work together to give much more positive results. In this study, the relationship between 3 determined terpene classes (myrcene, limonene and caryophyllene, terpenolene) and 11 different cannabinoid components (THCA, CBDA, A9-THC, CBD, THCV, CBDV, CBG, CBGA, CBC, A8-THC, CBN) was investigated. From machine learning problems, both regression and classification were performed. While 99.5% R2 value was reached with Linear Regression, 85.71% accuracy value was reached with Linear Discriminant Analysis. Terpene expression of plants with known cannabinoid information can be predicted with high accuracy. The results obtained are promising and guide agricultural engineers and institutions and organizations operating in the cannabis industry.