Predicting Stability and Antioxidant Activity of Apple Juice: A Machine Learning Regression Approach Using Pomegranate Peel Extract and Chitosan

Shanthi Vunguturi^* and Geeta Swarupa Pamidimalla

Department of Chemistry, Muffakham Jah College of Engineering and Technology, Hyderabad,Telangana, India.

Corresponding Author E-mail:India v.shanthi@mjcollege.ac.in

DOI : http://dx.doi.org/10.12944/CARJ.13.2.9

Article Publishing History

Received: 15 Apr 2025
Accepted: 16 May 2025
Published Online: 26 may 2025

Review Details

Reviewed by: Dr. Dilek Çavuşoğlu
Second Review by: Dr. Raksha Banka
Final Approval by: Dr. José Luis da Silva Nunes

Article Metrics

 

     PDF Downloads: 217

Abstract:

This study investigates the effects of pomegranate peel extract (PPE) and chitosan on the pH and antioxidant activity of apple juice, using a Random Forest Regression model for data analysis. Saccharomyces species, along with other yeasts, are the primary causes of spoilage in apple juice through fermentation. Natural preservatives, such as PPE and chitosan, are preferred over synthetic alternatives due to their antimicrobial, antioxidant, and antifungal properties. PPE, in particular, is rich in polyphenols that demonstrate strong antibacterial and antifungal activities, while chitosan is commonly used for its beneficial effects in food preservation and as a clarifying agent. The results indicate that PPE has a mild alkaline effect on the pH of apple juice, while the combination of PPE and chitosan leads to a more complex interaction that slightly decreases the pH. In terms of antioxidant activity, both PPE and chitosan enhance the juice’s antioxidant properties, with PPE being the primary contributor. The Random Forest model demonstrated strong predictive capability, achieving Mean Squared Error (MSE) values of 0.0018 for both pH and antioxidant activity with PPE alone, and 0.0083 when combined with chitosan. Compared to Support Vector Regression (MSE = 0.1450 for pH and 0.0178 for antioxidant activity), The Random Forest model effectively predicted both pH and antioxidant activity, with lower Mean Squared Error (MSE) values compared to the Support Vector Regression model, suggesting better performance. However, the model's negative R² scores highlight the need for further refinement, particularly in understanding the complex interactions between PPE and chitosan. The findings support the potential of PPE and chitosan as natural, safe alternatives to synthetic preservatives, offering a promising approach to enhancing the nutritional quality, stability, and shelf life of apple juice. Future research should focus on optimizing the concentrations of these additives and exploring their long-term effects, while also improving machine learning models to capture the intricate relationships in food preservation.

Keywords:

Antioxidant activity; Chitosan; MSE; PPE; Random Forest Model; Shelf life; Support Vector Regression Model

Download this article as: 

[ HTML Full Text]