Prediction of coliforms and Escherichia coli on tomato fruits and lettuce leaves after sanitizing by using Artificial Neural Networks

The objectives of this study were to investigate the efficacy of two sanitizers, i.e. hypochlorous and peracetic acids, in reducing coliforms and Escherichia coli levels on tomato fruits and lettuce leaves, and to mathematically predict the relationship among the initial bacterial load, type of vegetable/fruit, types and concentration of sanitizer and residual microorganism levels after the sanitizing, by applying artificial neural networks (ANNs). The E. coli and coliforms used in this study were isolated from the two food types, and their cultures were activated in Tryptic Soy Broth (ca. 6–7 log10 cfu/ml) before inoculating onto the fruit and vegetable. Both sanitizers reduced the number of the micro-organisms. However, as the hypochlorous acid concentration was increased, the level of viable coliforms and E. coli on the tomato fruits was reduced around 2–3 log10 cfu/g (p ≤ 0.05), compared to only about 1 log10 cfu/g reduction on lettuce leaves (p ≤ 0.05). Conversely, when the peracetic acid concentration was increased, the coliforms and E. coli levels on tomato fruits were reduced by some 3–4 log10 cfu/g (p > 0.05) compared to only about 2 log10 cfu/g on lettuce leaves (p > 0.05).The best sum square error from the neural prediction of residual coliforms and E. coli were 0.50 and 0.84, respectively, and the maximum R2 of residual coliforms and E. coli were 0.85 and 0.72, respectively. Only one hidden layer with three hidden neurons for coliforms and five for E. coli, were required to model this data.