Performance Analysis of Different Decision Tree Based Methods for Identifying Drop Out Students

The main concern of an educational institution is to enhance the performance of students. But dropping out of students is a large obstacle that creates bad impact in this purpose. For this reason, most of the educational institutions are conscious about reducing the rate of drop out students. To reduce the drop out students, firstly it is needed to identify the reasons behind the drop out students. In this paper to address the issue, different Decision Tree (DT) based methods as well as the boosted approaches have been used. Firstly, the features of the dataset have been sorted according to their importance based on Information Gain (IG). After the features being sorted, the dataset is partitioned into three categories: the all features, two-third of the features and one-third of the features. Thereafter the standard DTs and the boosted DTs are being experimented over the partitioned dataset to identify drop out students.