Development of a machine learning model for predicting Escherichia coli growth under different environmental conditions

Abstract

Escherichia coli (E. coli), a Gram-negative bacterium predominantly inhabiting the intestines of warm-blooded animals, including humans,  encompasses both benign and pathogenic strains. The ability of this strain to persist and proliferate in food matrices underscores the  critical importance of effective control measures and predictive tools in ensuring food safety across the food production and distribution  chain. This research investigated the development of an Artificial Neural Network (ANN) model for predicting Colony Forming Units (CFU)  of E. coli based on selected environmental factors such as temperature and pH. The study involved the collection of CFU data  under varying conditions, with temperatures ranging from 25 °C to 50 °C and pH levels from 2 to 12. The ANN model demonstrated a  high predictive accuracy, achieving an R-squared value of 98%, indicating strong correlations between predicted and actual CFU values.  The results showed that the optimal growth temperature for E. coli was 35 °C and pH of 7 (neutral), where the predicted CFU closely  matched the actual count. Additionally, the model proved effective across a range of conditions, confirming its reliability as a tool for  predicting microbial growth. These findings underscore the potential application of the ANN model in fields such as food safety, microbiology, and environmental monitoring, providing a valuable resource for controlling bacterial populations in various settings.