Comparison of Classification Data Mining C4.5 and Naïve Bayes Algorithms of EDM Dataset

• Published in: TEM Journal Vol. 10; no. 4; pp. 1738 - 1744
• Main Authors: Santoso, Joseph Teguh, Sri Rahayu Ginantra, Ni Luh Wiwik, Arif, Muhammad, Riinawati, R, Sudrajat, Dadang, Rahim, Robbi
• Format: Journal Article
• Language: English
• Published: Novi Pazar UIKTEN - Association for Information Communication Technology Education and Science 01.11.2021
• Subjects: Accuracy; Algorithms; Boarding schools; Breast cancer; Classification; Data mining; Datasets; Decision trees; ICT Information and Communications Technologies; Methods; Performance evaluation; Students; Naive Bayes; Comparison; C45; data mining; Classification; Performance; EDM
• ISSN: 2217-8309; 2217-8333; 2217-8333
• DOI: 10.18421/TEM104-34

The purpose of this research is to choose the best method by comparing two classification methods of data mining C4.5 and Naïve Bayes on Educational Data Mining, in which the data used is student graduation data consisting of 79 records. Both methods are tested for validation with 10-ford X Validation and perform a T-Test difference test to produce a table that contains the best method ranking. Different results were obtained for each method. Based on the results of these two methods, it is very influential on the dataset and the value of the area under curve in the Naïve Bayes method is better than the C4.5 method in various datasets. Comparison of the method with the 10-Ford X Validation test and the T-Test difference test is that the Naïve Bayes method is better than C4.5 with an average accuracy value of 73.41% and an under-curve area of 0.664.