Gas sensor-based machine learning approaches for characterizing tarragon aroma and essential oil under various drying conditions

Aroma is one of the most significant quality traits for many pharmaceutical plants and their products as it indicates the quality of the raw material. The aroma may be lowered to imperceptible levels, altered, or damaged during the processing of herbs, such as drying or fermentation. Using an electronic nose (e-nose) is one of the most efficient approaches for identifying and evaluating the aroma of essential oils (EO). In this study, tarragon was dried in a hybrid dryer designed explicitly for frying the herbs at four temperatures (40, 50, 60, and 70 °C) and three air velocities (1, 1.5, and 2 m/s). After extracting its EO, the purity of the tarragon EO was assessed by an e-nose comprising nine metal oxide semiconductor (MOS) sensors. The highest EO levels (0.359) was obtained upon drying at 40 °C. By increasing the temperature from 40° to 70°C, the EO was declined, and its lowest level (0.26) was assessed at 70 °C. multivariate data analysis and artificial neural networks modeling were also employed to quantify and classify the obtained EOs based on the output of the sensor. Multivariate discrimination analysis (MDA) and Quadratic discriminant analysis (QDA) offered a 100 % accuracy in classifying 12 groups of EO.