M.A. Thesis:

Comparison of the functions of sigmoid logarithm and hyperbolic tangent in modeling the root extraction of Morinda citrifolia plant in soybean oil stability with the help of ultrasonic pretreatment

Due to the various adverse effects of synthetic antioxidants, the utilization of plant antioxidants has received much attention in recent years to delay or prevent food oxidation, especially foods based on oil or fat. This study examines the antioxidant compounds and the volume of folin compounds in Morinda Citrifolia (Noni) extract by the extraction method with the help of ultrasonic waves. The results of comparing the neural network of the genetic algorithm with the response surface methodology showed a better ability of this method. Optimal neural network includes 10 hidden neurons in the first layer and 10 hidden neurons in the second layer. This network architecture had the ability to predict the efficiency of extracting antioxidant compounds with a high correlation coefficient equal to 0.99. In the final step of the study, the optimal sample of Morinda Citrifolia root extract was selected by neural network modeling and soybean oil was added to the root extract of Morinda Citrifolia by maceration and synthetic antioxidant (BHT) method. The results obtained from oxidative stability of the oil illustrated that the optimal sample was more effective in reducing peroxide index and thiobarbiotic acid (TBA) than the control sample and synthetic antioxidant (BHT). Keywords: Morinda Citrifolia (Noni) Extract, Extraction, Soybean Oil, Antioxidant, Thermal Stability.

Keywords: Keywords: Morinda Citrifolia (Noni) Extract, Extraction, Soybean Oil, Antioxidant, Thermal Stability