Finally, in the last phase (Step 5), the data was analyzed by each of the models, and the best one was chosen. In other words – the model that best predicted the class attribute value, with a high accuracy rate was selected. (Each model was evaluated on a training set and its accuracy was evaluated again on the test set respectively). As much as the ‘gap’ between the trained predicted value and the actual data was lower the model was found as a better predictor for analysis regarding the actual data.