Models for Mold Infection and Mycotoxin Production and Influencing Factors: A Review

Mold infection and mycotoxin production, driven by fungi such as Aspergillus, Fusarium, and Penicillium, pose significant threats to global food safety, contributing to 25% of crop losses annually [Eskola et al., 2020]. This review synthesizes mathematical modeling approaches—empirical, mechanistic, and artificial intelligence (AI)-based—for predicting mold growth and mycotoxin contamination in food systems. Empirical models, like polynomial regressions, offer simplicity but limited generalizability, while mechanistic models, such as the Baranyi-Roberts framework, provide biological insights yet demand detailed data. AI-driven models, including deep learning, achieve up to 95% predictive accuracy by capturing nonlinear environmental interactions (e.g., temperature, water activity) [Mateo et al., 2021]. Key factors influencing contamination—temperature, moisture, pH, oxygen, and substrate—are analyzed, with AI enhancing real-time risk assessment. Challenges include data scarcity, model interpretability, and high costs, particularly in developing regions like Vietnam. By integrating hybrid AI-mechanistic models and leveraging IoT for real-time monitoring, future strategies can reduce mycotoxin risks, supporting safer storage and sustainable food systems. This review guides researchers and policymakers in advancing predictive tools for food safety management.

Keywords: Artificial Intelligence (AI), Food Safety, Mold Infection, Mycotoxins, Predictive Modeling.