Estimation of peroxidase activity in red cabbage by artificial neural network

Enzymes in plant tissues can have undesirable or desirable effects on the quality of fruits and vegetables such as the post-harvest senescence, oxidation of phenolic substances, starch-sugar conversion and post-harvest demethylation of pectic substances leading to softening of plant tissues during ripening. Peroxidase (POD) is a commonly enzyme in vegetable, with bad effects on quality of their products. Structure of POD and POD isoform differ in each vegetable. Hence, activity of POD is unique to each vegetable. The aim of this study was evaluation efficiency of different essential oils as natural antioxidant in POD inactivation and estimation of POD activity by artificial neural network (ANN) modeling. In this study we used natural antioxidant (cumin, fennel, clove) in red cabbage. An ANN was developed by using a multilayer perceptron model, three input neurons (type and concentration of antioxidant and duration of enzyme activity), one hidden layer and 21 hidden neurons. The ANN model predicted POD activity with a mean square error of 0.0002629 and a good correlation between predicted and experimental data (R2=0.9974). These results show the ability of ANN technology for predicting POD activity of red cabbage under natural antioxidants.