Algorithm Implementations Naïve Bayes, Random Forest. C4.5 on Online Gaming for Learning Achievement Predictions

Abstract

The online game is a game which is currently booming and interest ranging from children, teens, to adults. Online games can create a sense of opium to the people who play it. Online games become a new problem for the students, because online games make learning impaired concentration. The learning achievements can be measured from the value of report cards. The challenge on this research can be carried out using a method of classification for predicting learning achievements using algorithms of classification i.e. Naïve Bayes, Random Forest, and C4.5. After the third comparison algorithm, then the prediction results obtained by learning achievements. Naïve Bayes algorithm proved that value the accuracy and value of the AUC 69.18% of 0.771 contains the classification, fair for the random forest algorithm accuracy 66.34% and AUC values of 0.738 contains the classification, fair as for algorithm C4.5 65.65% accuracy and value of the AUC of 0.686 including into poor classification. From these results it can be concluded that the naïve bayes algorithm has higher accuracy compared with the random forest algorithm and C4.5, visible difference in accuracy between the naïve bayes with random forest of 2,84%, whereas the difference between the naïve bayes with C4.5 of 3,53%. Naïve bayes algorithm is thus able to predict achievement students can study better. Keywords—online games; learning achievement; naïve bayes algorithm; random forest; C4.5