Modeling the growth of Salmonella typhimurium under the effect of Zataria multiflora essential oil, pH, and temperature by artificial neural networks

Salmonella is the most important cause of bacterial food-borne disease outbreaks in the world. In this study, we have trained and validated artificial neural networks (ANNs) to predict the combined effect of Zataria multiflora essential oil (EO), pH, and temperature on the probability percentage of growth initiation (log P%) of Salmonella. Z. multiflora was collected in the Fars province of Iran. Lyophilized cultures of Salmonella typhimurium ATCC 25923 was used in this study. This design included four levels of EO (0.0, 0.015, 0.03, and 0.06%), three levels of pH (5.5,6, and 7.3), three storage temperatures (35, 25, and 15°C), and repeated observations (18 times) for growth in brain heart infusion broth for up to 43 days. We have designed a standard and the so-called feed-forward ANN, including four input neurons, eight neuron in hidden layer, and one output neuron to predict the combined effect of Z. multiflora EO, pH, and temperature on the probability percentage of growth initiation (log P%) of S.typhimurium. The mean and standard deviation of ANN and real outputs were −2.9771 ± 2.43 and −2.9722 ± 2.39, respectively. The mean differences (and 95% CIs) between the ANN and real outputs were 0.0049 (0.0009–0.0089). Result showed better prediction compare to the previous study(R = 0.998).