Akademik Veri Yönetim Sistemi
APPLIED FOOD RESEARCH, cilt.5, sa.2, 2025 (ESCI)
Food contact surfaces (FCS) play a crucial role in food processing environments, and maintaining their hygiene is essential for preventing foodborne illnesses. Rhamnolipids, biodegradable and eco-friendly biosurfactants, have gained interest in the food industry for their sustainable antimicrobial properties. In this study, L. monocytogenes demonstrated the highest sensitivity to rhamnolipid treatment, with the lowest minimum inhibitory (1.5625 mu g/ mL) and bactericidal (3.125 mu g/mL) concentrations, compared to Salmonella typhimurium and Escherichia coli. Due to its increased susceptibility, L. monocytogenes was further used to assess the decontamination efficacy of rhamnolipids on stainless steel, wood, and HDPE surfaces, each inoculated to achieve consistent microbial loads. When applied at its MBC level, rhamnolipid exhibited rapid bactericidal action, eliminating L. monocytogenes within 10 min on stainless steel and wood. However, its effectiveness was diminished on HDPE, where high bacterial counts (4.9 log10 CFU) persisted after 40 min. To evaluate bacterial count dynamics, three mixed-effects models, Generalized Estimating Equation (GEE) models, and fixed-effects (OLS) models were developed. We evaluated multiple models, including Random Forest, Gradient Boosting, Bagging, AdaBoost, Linear Regression, Elastic Net, and SVR, under two cross-validation strategies: 5-Fold CV and Leave-One-Out CV (LOOCV). Among all models, Gradient Boosting (LOOCV CV-MSE: 0.98, Test RMSE: 1.30) and Random Forest (LOOCV CV-MSE: 1.10, Test RMSE: 1.18) demonstrated the best predictive performance. Feature importance analysis using Random Forest revealed treatment type (importance = 0.48) and incubation time (importance = 0.41) as the most critical predictors. These results underscore the importance of rhamnolipids for FCS decontamination and highlight the predictive power of machine learning in food safety applications.