The effect of infrared on Bacillus cereus in paprika powder: Modeling through genetic algorithm-artificial neural network

In this study, the effect of infrared (IR) on decontamination of Bacillus cereus, color, weight losses, and temperature profiles at paprika powder was determined in difference IR radiation power (100, 200, and 300 W), different sample distances from a radiation source (5, 10, and 15 cm), and various holding times. The most reduction of B. cereus count (2.3 log CFU/g) was achieved after 1 min holding time at 200 W IR power and 5 cm distance. The highest D-value (0.18 min) was achieved after a holding time of 0.5 min at 300 W IR power and 5 cm distance. The a* value of paprika powder was slightly affected and the highest color change was observed at 100 W IR power, 10 cm distance, and 8 min resulting in a decrease of a* from 42.537 ± 0.201 to 38.645 ± 0.429. Data were analyzed to predict the antibacterial effects of IR on B. cereus in paprika powder through an artificial neural network (ANN) model. The developed GA-ANN, which included 20 hidden neurons, could predict the B. cereus population with R2 = 0.9561. The results indicated that the GA-ANN model could give a good prediction for the population of B. cereus. Sensitivity analysis results showed that IR irradiation time was the most sensitive factor for the prediction of the B. cereus population.