The model is adjusted, or trained, using a collection of data from a given source as input, typically referred to as the training set. After successful training, the neural network will be capable to perform classification, estimation, prediction, or simulation on new data from the same or similar sources [16].