Discovering the Chemical Factors Behind Regional Royal Jelly Differences Via Machine Learning

This study explored how key chemical parameters could be used to distinguish royal jelly (RJ) from different regions of Turkey using a data-driven machine‐learning approach. A total of 84 RJ samples from 13 distinct geographical locations were analysed for four chemical characteristics: moisture (%), pH, acidity (meq/kg), and 10-hydroxy-2-decenoic acid (10-HDA %). Statistical analysis (ANOVA) revealed significant regional differences among the samples. The descriptive mean values were: moisture ~ 63.05 % ± 2.99, pH ~ 3.67 ± 0.08, acidity ~ 45.32 ± 3.55 meq/kg and 10-HDA ~ 2.40 ± 0.24 %. The machine-learning model (decision-tree algorithm) indicated that the 10-HDA content was the most prominent single predictor for region classification of the RJ samples. Thus, the authors suggest that 10-HDA may serve as a key marker for regional authentication of Turkish royal jelly.