Comparative Machine Learning Modeling of Infrared Drying Kinetics in Cactus Fruit (Opuntia ficus-indica) Slices

The objective of this study was to model the infrared drying kinetics of cactus fruit (Opuntia ficus-indica) slices using advanced machine learning (ML) approaches. Drying experiments were conducted at a constant temperature of 70 °C using slice thicknesses of 2, 5, and 8 mm. Approximately 200 experimental data points describing the temporal evolution of moisture ratio (MR) were obtained. In previous analyses, the Midilli–Küçük model was identified as the most suitable semi-empirical thin-layer model for this dataset. In the present study, the same experimental data were re-evaluated using nonlinear ML algorithms to further improve predictive accuracy. Support vector machines (SVM), artificial neural networks (ANN), random forest (RF), and linear regression (LR) were employed. Drying time and slice thickness were used as input variables, while moisture ratio was defined as the output variable. Model performance was evaluated using a rigorous 10-fold cross-validation procedure. The results indicated that the SVM model achieved the highest prediction accuracy, with a coefficient of determination of R² ≈ 0.9998 and a root mean square error of approximately 0.005, followed closely by the ANN model (R² ≈ 0.9990). In contrast, the linear regression model failed to adequately capture the nonlinear characteristics of the drying process. Overall, the findings demonstrate that SVM and ANN provide robust and accurate alternatives to conventional empirical thin-layer models for predicting infrared drying kinetics of cactus fruit.