Finally the best model was chosen and the results were evaluated. By comparison the results of prediction for student’s during the year 2012/13 according to the prediction class attribute: ‘‘Final grade’’, with the actual Final grade of the same student the authors could choose the best parameters for the Fill from example data mining technique. The final Data mining Fill from example predictive model was ready for prediction as a ‘future student Final grade ’, which was based on already known instances. The model could be used by confidence, until several assumptions were obtained as True (Berry & Linoff, 2000). The first assumption assumed that the past results were considered as good predictors of the future (if the student generation is changed, therefore; their attitude and behavior may vary from the past and create different student pattern). The next assumption assumed that the data was available at hand. The researchers assumed that the data will be always available for HEIs. The last assumption assumed that the data contained what the researchers aimed to predict – the research data mining model as a result contains the relevant data for analysis.