Applying an intelligent model and sensitivity analysis to inspect mass transfer kinetics, shrinkage and crust color changes of deep-fat fried ostrich meat cubes

- IA and ANN could be employed for online prediction of some physicochemical properties.
- GFF networks with two hidden layers were found the best models.
- Mass transfer kinetics were so sensitive to frying temperature.
- Microwave power was the most important parameter to changes of color and shrinkage.

The objectives of this study were to use image analysis and artificial neural network (ANN) to predict mass transfer kinetics as well as color changes and shrinkage of deep-fat fried ostrich meat cubes. Two generalized feedforward networks were separately developed by using the operation conditions as inputs. Results based on the highest numerical quantities of the correlation coefficients between the experimental versus predicted values, showed proper fitting. Sensitivity analysis results of selected ANNs showed that among the input variables, frying temperature was the most sensitive to moisture content (MC) and fat content (FC) compared to other variables. Sensitivity analysis results of selected ANNs showed that MC and FC were the most sensitive to frying temperature compared to other input variables. Similarly, for the second ANN architecture, microwave power density was the most impressive variable having the maximum influence on both shrinkage percentage and color changes.