Vol. 2, Issue 2, Part A (2025)

Fungal chitosan, derived from agro-waste through fermentation, has gained significant attention due to its versatile applications in biomedicine, agriculture, and food industries. Optimizing fermentation yields is crucial for large-scale production, which is often influenced by various agro-waste input variables. Machine learning (ML) offers a promising approach to predict and optimize fermentation yields by analyzing complex relationships between these variables. This research aims to develop a predictive modeling framework using ML techniques to enhance fungal chitosan production. Data from fermentation experiments with various agro-waste inputs were used to train and validate multiple ML models. The input variables considered include carbon and nitrogen sources, pH, temperature, and incubation time, which are known to impact the fermentation process. Several algorithms, including regression analysis, random forests, and support vector machines, were evaluated for their predictive accuracy and ability to generalize across different fermentation setups. The model's performance was assessed using root mean square error (RMSE) and R-squared values. The results indicated that ML models could effectively predict the fermentation yield, providing valuable insights for optimizing production parameters. Moreover, the integration of ML-based prediction systems could lead to cost-effective and scalable fungal chitosan production. The research highlights the potential of ML in the biotechnological field and offers a framework for improving the efficiency of fermentation processes in agro-industrial waste valorization.