Volume 22, Issue 161 (2025)
FSCT 2025, 22(161): 121-136 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Maroufpour B, Ziaiifar A M, Sabbaghi H, Ghorbani M, Yalghi S. Prediction of moisture and density in shrimp during hot air frying with artificial neural networks model. FSCT 2025; 22 (161) :121-136
URL: http://fsct.modares.ac.ir/article-7-76585-en.html
URL: http://fsct.modares.ac.ir/article-7-76585-en.html
Bahareh Maroufpour1 
, Aman Mohammad Ziaiifar * 2, Hassan Sabbaghi3 
, Mohammad Ghorbani1 , Saeed Yalghi4
, Aman Mohammad Ziaiifar * 2, Hassan Sabbaghi3 , Mohammad Ghorbani1 , Saeed Yalghi4
1- Gorgan University of Agricultural Sciences and Natural Resources
2- Gorgan University of Agricultural Sciences and Natural Resources ,ziaiifar@gau.ac.ir
3- Assistant Professor, Department of Food Science and Technology, Faculty of Agriculture and Animal Science, University of Torbat-e Jam, Torbat-e Jam, Razavi Khorasan Province, Iran
4- Golestan Province Fisheries Research Center
2- Gorgan University of Agricultural Sciences and Natural Resources ,
3- Assistant Professor, Department of Food Science and Technology, Faculty of Agriculture and Animal Science, University of Torbat-e Jam, Torbat-e Jam, Razavi Khorasan Province, Iran
4- Golestan Province Fisheries Research Center
Abstract: (113 Views)
In this research, artificial neural networks (ANN) was presented to predict changes in moisture and density of shrimp during hot air frying process (at three temperatures of 140, 160 and 180 degrees Celsius for 15 minutes). Neural networks in the form of multilayer perceptron (MLP) with sigmoid tangent transfer function in the hidden layer and linear transfer function in the output layer was designed to predict moisture (with two inputs: temperature and time) and density (with three inputs: temperature, time and moisture) in MATLAB software. Different backpropagation algorithms include Levenberg-Marquardt, Gradient descent, Gradient descent with adaptive learning rate, Adaptive learning rate backpropagation, Gradient descent with momentum, and Scaled conjugate gradient. The structure of the models was validated by calculating the coefficient of determination (R2), root mean square error (RMSE) and mean absolute error (MAE). Finally, the importance of the inputs in terms of the effect on the output variable for predicting moisture and density was investigated by designing the default hyperbolic tangent neural networks in SPSS software. The results showed that with the decrease in moisture and the development of pores in shrimp, the density of the product gradually decreased during hot air frying, and with the increase in the temperature of the process, a further decrease in density was observed. In the moisture model, the backpropagation algorithm of Gradient descent with momentum (R2=0.989, RMSE=0.171, MAE=0.131) and in the density model, the Levenberg-Marquardt algorithm (R2=0.974, RMSE=0.0096, MAE=0.0066) showed the minimum error in testing. In the moisture neurocomputing, the importance of time and temperature variable was equal to 0.883 and 0.117, respectively. In the density neurocomputing, the importance of moisture, time and temperature variables were 0.588, 0.278 and 0.134, respectively. The Findings can be used in the design of artificial intelligence for controlling and creating automation in hot air fryers.
Article Type: Original Research |
Subject:
Food Chemistry
Received: 2024/08/14 | Accepted: 2024/10/7 | Published: 2025/06/22
Received: 2024/08/14 | Accepted: 2024/10/7 | Published: 2025/06/22
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |