Comparison of naive Bayes classifier and C4.5 algorithms in predicting student study period

Abstract

This study aims to compare Naive Bayes Classifier algorithm with C4.5 in predicting the study period of student. The data for this study is collected from Department of Informatics UIN Sunan Gunung Djati Bandung. Software development method that used is Rational Unified Process (RUP). Comparison of both algorithms is the accuracy of predicting student study period and the speed of data processing. The attributes that used are student id, name, gender, GPA, entry point, tahfidz (cited the holy Quran as a special attribute on the case study), previous school, and extra activities during lecture. The result of this study shows that Naive Bayes Classifier and C4.5 algorithm can be applied to predict the study period of students well and accurate enough. The accuracy of Naive Bayes Classifier algorithm is around 88% better slightly than the C4.5 algorithm that has accuracy around 87%. However, the processing time of algorithm C4.5 is better than the algorithm Naive Bayes Classifier, C4.5 has processing time 0.003 nanosecond and the Naive Bayes Classifier has 12.7 nanosecond.