Decision Tree Algorithm to Improve the Learning Discipline Classification Model of Group Guidance Students at MTs Darul Mutta’alimin

Abstract

This study aims to find patterns of student learning discipline at MTS Darul Mutta'alimin. In addition, this research also seeks to identify the main factors that influence discipline and build an analysis model that can be used by teachers and mentors to effectively improve student discipline. In its implementation, this research uses the Knowledge Discovery in Databases (KDD) method which consists of several stages, namely data collection, preprocessing, transformation, data mining, and evaluation of results. The data used includes test scores, student participation in class, and various other behaviors related to discipline. The results showed that the Decision Tree model developed had a high level of accuracy, reaching 98.36%. The main factors found to influence discipline are “Test Grades” and “Class Participation”. In the process of classifying students into disciplined or undisciplined categories, the confusion matrix shows excellent model performance with a very low error rate. These results can be applied in education to assist teachers in monitoring students who need special attention. This study proves that the Decision Tree algorithm is very effective for finding patterns in data. For further development, it is recommended to add non-academic factors and test other algorithms to improve accuracy and broader generalization of the model