NEURAL NETWORK MODELING OF OIL YIELD FROM SHEAKERNELS IN A HYDRAULIC PRESS
An investigation on prediction of oil from shea kernels in a hydraulic press subject to process variables such as moisture content, pressing time, applied pressure, heating time and heating temperature was carried out. Artificial neural network (ANN) technique was applied using experimental data from a previous study. These data were then used for network training and testing. The back propagation technique was then used for establishing the network. The prediction accuracy of the neural network model was significantly improved compared to statistical model. (R=0.96)
Key words: Oil expression, yield, neural network, prediction.