Prediction of the moisture ratio of Atama (Heinsia Crinita) leaves using artificial neural network (ANN)

In this work, an artificial neural network (ANN) model was used to predict the moisture ratio of atama (Heinsia crinita) dried under different drying temperatures of 40 0C, 50 0C, 60 0C, and 70 0C using a laboratory dry oven. The experimental data collected (140 data points in all) which was partitioned into three sets: training (70%), validation (15%), and testing (15%) were modeled using artificial neural network (ANN), an Artificial Intelligence approach. The ANN model architecture of ANN (3 – 4 - 1) used in this work was selected by trial-and-error approach. The input layer had three (3) inputs (drying rate, temperature, time), the hidden layer had four (4) neurons, and the output layer had one (1) output (moisture ratio). Levenberg-Marquardt (LM) algorithm was used for training the network, and TANSIG and Purelin transfer/activation functions were used for the hidden layer and output layer, respectively. The model had a learning rate of 0.7, and the number of epochs was set at 1000. The results obtained showed that the ANN methodology could precisely predict experimental data with high correlation coefficient (R-Squared) value of 0.9995 – 0.9977 and low mean square error (RMSE) of 0.00052568, as the artificial neural network model more accurately predict the drying parameter (moisture ratio). The sensitivity analysis performed shows that temperature has the greatest impact on the moisture ratio of atama. From the finding, the ANN technology which is embedded in the neural toolbox of MATLAB mathematical software is indeed a tool of choice when it comes to the prediction of parameters of non-linear and complex processes like drying. The unique modelling technique and the model it evolved represent a huge step in the trajectory of achieving full automation of moisture ratio estimation which will increase the utilization of atama as well as other vegetables to curb the unending events of food spoilage currently plaguing the global food and agriculture industry.