The developing of prediction model Premier League football results with datamining technique

Abstract

The purposes of the research were to 1) study and development of prediction model of Premier League football results with a data mining technique, 2) assess the efficiency of prediction and planning for decision support Premier League football. The participants all were by purposive samplings and 3992 records. The process of research CRISP-DM is six steps and two technique data models in the Decision Tree and Naïve Bayes. The statically is percentage and accuracy.

           The research findings showed that the developing of a prediction model for Premier League football results in a Decision Tree which a high factor of date of more than 77.80, with the referee football name being M Dean and R East on a date less than equal to 77.80 the result Half Time Away Team Goals (HTAG) then if PSCA team more than 1.72 to result in Home Team Fouls Committed (HF) and If less than or equal to 1.72, results in a draw 2) the performance of the model percentage accuracy of 2 techniques: Decision Tree, Naïve Bayes. The average accuracy was 80.67% at a high level. Furthermore, the technique with the highest accuracy was Decision Tree at 98.58%, followed by the Naïve Bayes technique at 62.75%.