This study proposed a novel approach for modeling and predicting the moisture ratio of yoghurt powder during microwave‐assisted foam‐mat drying using artificial neural networks (ANN) and multiple linear regression (MLR). Various ANN configurations, including different training algorithms, transfer functions, and hidden‐layer neuron numbers, were systematically evaluated using the coefficient of determination ( R ² ) and the root mean squared error (RMSE) to find the best structure. In addition, three stepwise MLR models with different predictor combinations were developed to assess the contribution of the independent variables to moisture ratio prediction. The optimal ANN model, consisting of 15 hidden neurons with the logsig activation function in the hidden layer and the purelin activation function in the output layer, trained using the Levenberg–Marquardt backpropagation algorithm ( trainlm ), achieved the highest predictive performance with R ² of 0.9979 and RMSE of 0.0296. Garson's algorithm was employed to assess the relative importance of the input variables (drying time, egg albumin concentration, and microwave power) based on the optimal ANN model. The sensitivity analysis indicated that drying time (40.86%) and egg albumin concentration (40.41%) were the most influential factors governing moisture ratio prediction. Among the multiple linear regression models, MLR1 exhibited the best predictive performance. Nevertheless, the ANN model outperformed all MLR models by achieving higher prediction accuracy and better convergence. These findings demonstrate that the developed ANN model is a reliable tool for predicting moisture ratio and has considerable potential for process optimization, operational control, and reducing the need for extensive experimental trials.

The purpose of the study was to assess the prediction ability of multiple linear regression (MLR) and artificial neural network (ANN) models for the browning index, total carotenoid content, 5-hydroxymethylfurfural (HMF) and ascorbic acid of orange juice during storage after heat processing. For ANN models, the effect of neuron number of the hidden layer, epoch number, training algorithms and transfer functions are investigated using the trial-error method for selecting best design ANN models. The different methods (stepwise and enter) in multiple linear regression models were performed for detecting impact of independent variables on dependent variables. The performance of ANN and MLR models was determined through unseen data by means of statistical analysis. Regarding performance indices of ANN models for test data, overall R and R2 were recorded as follows: 0.92 and 0.84 (browning index), 0.99 and 0.98 (HMF), 0.92 and 0.86 (ascorbic acid), 0.97 and 0.94 (total carotenoid content), respectively. R and R2 values of MLR models for test data were 0.79 and 0.68 (browning index), 0.94 and 0.88 (HMF), 0.92 and 0.85 (ascorbic acid), 0.93 and 0.90 (total carotenoid content), respectively. Both models provided accurate predictions. However, the superior predictive power of ANN models is that they can learn directly from examples without calculating the parameters using statistical techniques. The results revealed that ANN models showed greater potential with high R and R2 value, and the lowest error values when compared with the MLR model, but both ANN and MLR had almost same performance for prediction of ascorbic acid.

The drying process is complex because of mass and heat changes. Many factors influence the drying procedure. The present work aimed to evaluate the possibility of artificial neural networks (ANN) and multiple linear regression (MLR) to characterize the moisture ratio of red beetroot slices during ultrasound assisted vacuum drying for multiple inputs. For models, inputs were temperature (T, 50o C), time (t, 0-900 min), weight (w, 13.01±0.25 g), pressure (p, 0.05 MPa) and pretreatment (PR: 0/1) and the moisture ratio (MR) was the output. The ANN models was obtained due to impact of the hidden layer's neuron number and size. MLR model was obtained with stepwise method. As performance criteria of all models, R2 , RMSE, and MAPE values were taken into account. The best ANN model was obtained by three layers (5 inputs, 15 neurons in one hidden layer and 1 output) with RMSE of 0.0117, MAPE of 2.293% and R2 of 0.9996 for all data. The results showed that the ANN model including temperature, time weight, pressure and pretreatment) is a more effective predictive tool since it can yield better outcomes than stepwise MLR model including weight, pressure and pretreatment. The applicability of models to various drying circumstances is restricted because the trials were carried out at a one temperature and vacuum pressure (50 °C and 0.05 MPa). Models with more capacity for generalization should be created by experimenting with various pressure and temperature circumstances. Models can be also enhanced for improved evaluation by utilizing various input combinations.