Modeling of antibacterial activity of annatto dye on Escherichia coli in mayonnaise

Abstract

Annatto ranks second in economic importance worldwide among all natural colorants and its extract fraught with antimicrobial and antioxidant properties. In the present paper, adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm–artificial neural network (GA–ANN) models were undertaken to predict the annatto dye on Escherichia coli in mayonnaise. The ANFIS and GA–ANN were fed with 3 inputs of annatto dye concentration (0%, 0.1%, 0.2% and 0.4%), storage temperature (4 and 25 °C) and storage time (1–17 days) for prediction of E. coli population. Both models were trained with experimental data. The results revealed that the annatto dye was able to decline E. coli and the bactericidal effect of annatto dye was stronger at 25 °C than that in 4 °C. The developed GA–ANN, included 13 hidden neurons, could predict E. coli population with coefficient of determination of 0.995. The largely agreement between experimental and ANFIS predictions data was also acceptable (R2=0.991). Sensitivity analysis results revealed that storage time was the most sensitive factor for prediction of E. coli population.