Rose geranium (Pelargonium graveolens L’Hér. var. Bourbon) was grown hydroponically in a 72 m² plastic-covered greenhouse tunnel during separate growing seasons, to explore the feasibility of predicting changes in rose geranium essential oil biosynthesis using data mining algorithms and MANOVA. Although the study was based on limited data during the separate seasons, the Pearson’s correlation coefficient and ANOVA showed a trend, where the higher essential oil yield (OY) and quality (C:G ratio) in rose geranium were possibly associated with the shorter, bushier plants, as well as the smaller leaf area. On the other hand, these yield and yield traits were observed to possibly affect foliar fresh mass (FFM), with an estimated FFM exceeding 1,129 g, which could yield approximately 2.7 g of essential oil. Interestingly, the PCA also demonstrated that the C:G ratio could be associated with the OY and FFM, corroborating the possible use of the models in future studies to explore predictions of essential oil biosynthesis under various growing conditions. Future research should focus on increasing true biological replications and further exploring data mining techniques, particularly the MARS and CART algorithms, to enhance predictive modelling of rose geranium essential oil yield and quality.
Khetsha et al. (2026) studied this question.