Evaluation of Carrot Qualitative Properties in Microwave Drying with Different Pretreatments by Artificial Neural Network

In this research, the carrot was first grounded in equal pieces; the pieces underwent pretreatment using Ohmic and blanching methods. The treatments were preheated at 3 ohmic voltage levels as 40, 60 and 80 Volt and 3 period levels of 2, 4 and 6 minutes. Then parameters examined in the pretreatment blanching in 3 period levels of 2, 4 and 6 minutes. After the pre-treatment, samples were dried in microwave in 3 power levels of 360, 600 and 900 W. The qualitative properties of samples including antioxidant, flavonoids and total phenol content were determined for pre-treated and dried samples. A multilayer neural network of perceptron was used to teach data using 3 and 6 neurons in the hidden layer and three functions of hyperbolic, sigmoid, and linear tangent activation. According to the results obtained from the network, for Ohmic premixing method the best values of R2 and RMSE were observed for antioxidant content (sigmoid-6 neuron), flavonoids content (hyperbolic tangent-6 neuron) and total phenol content (sigmoid-6 neuron). Also, the activation network with 3 neurons in the hidden layer and the hyperbolic tangent activation function were able to obtain the best values of R2 and RMSE for the total phenol content, flavonoids and antioxidants in the blanching pre-treatment.