Volatile organic compounds (VOCs) are central to the aromatic and therapeutic properties of essential oils (EOs), with their profiles serving as reliable indicators of EO quality. In Cistus ladanifer, the synthesis and emission of VOCs—particularly terpenic hydrocarbons—are strongly influenced by environmental variables such as temperature, humidity, and solar radiation. Traditional EO quality assessment methods, including gas chromatography (GC), although highly accurate, are costly, labor-intensive, and destructive. This study proposes a smart sensor system that utilizes a low-cost array of MQ gas sensors combined with machine learning (ML) algorithms for non-destructive classification of VOC fingerprints in Cistus ladanifer EO. VOC data from 33 EO samples were collected using MQ sensors and paired with environmental datasets (Daily, 15-Day, and All) obtained from weather stations in the cultivation areas. Gas chromatography coupled with flame ionization detection and mass spectrometry (GC-FID/MS) was used as the reference method to quantify the concentrations of terpenic hydrocarbons. A total of 5,154 data points were used, with 75% used for training and validation and 25% reserved for independent testing. Model performance was evaluated using tenfold cross-validation within the training set. Several ML models—Neural Networks, Decision Trees, Logistic and Linear Regression, and Ensemble—were trained to predict hydrocarbon levels from VOC signals and environmental inputs. The “Daily” dataset produced the highest R2 (92.68%), followed closely by the “All” dataset (R2 = 92.67%). Minimal feature subsets were identified to optimize model performance: 8, 5, and 13 features for the Daily, 15-Day, and All datasets, respectively, using environmental variables alone. When VOC data were integrated, optimal feature sets increased to 12, 20, and 29. This study demonstrates the feasibility of a smart sensor-based framework for EO quality assessment via gas fingerprint classification. The integration of low-cost sensing technologies with ML models enables a scalable, rapid, and non-invasive approach to monitoring EO composition. While results were obtained under controlled laboratory conditions and real-world performance will require field validation to address environmental variability and sensor drift, the framework shows strong potential for deployment in precision agriculture in coastal areas and beyond, environmental monitoring, and the sustainable valorization of aromatic plant resources.
Ahmad et al. (Sat,) studied this question.