Predicting Students Drop Out a Case Study

Only available on StudyMode
  • Download(s) : 51
  • Published : September 27, 2010
Open Document
Text Preview
Predicting Students Drop Out: A Case Study
Gerben W. Dekker1, Mykola Pechenizkiy2 and Jan M. Vleeshouwers1 g.w.dekker@student.tue.nl, {m.pechenizkiy, j.m.vleeshouwers}@tue.nl 1 Department of Electrical Engineering, Eindhoven University of Technology, the Netherlands 2 Department of Computer Science, Eindhoven University of Technology, the Netherlands Abstract. The monitoring and support of university freshmen is considered very important at many educational institutions. In this paper we describe the results of the educational data mining case study aimed at predicting the Electrical Engineering (EE) students drop out after the first semester of their studies or even before they enter the study program as well as identifying success-factors specific to the EE program. Our experimental results show that rather simple and intuitive classifiers (decision trees) give a useful result with accuracies between 75 and 80%. Besides, we demonstrate the usefulness of cost-sensitive learning and thorough analysis of misclassifications, and show a few ways of further prediction improvement without having to collect additional data about the students.

1

Introduction

The monitoring and support of the first year students is a topic that is considered very important at many educational institutions. At some of the faculties yearly student enrollment for a bachelor program can be lower than desired, and when coupled with a high drop out rate of freshmen the need in effective approaches for predicting student drop out as well as identifying the factors affecting it speaks for itself. At the Electrical Engineering (EE) department of Eindhoven University of Technology (TU/e), the drop out rate of freshmen is about 40%. Apart from the department’s aim to enforce an upper bound to the drop-out rate, there are other reasons to want to identify successful and unsuccessful students in an early stage. In the Netherlands, there is the legal obligation that universities have to provide students with the necessary support to evaluate their study choice. In general, students who choose to pursue their study career at another institution, should do this at an early stage. For EE students there is a very concrete reason to evaluate before the end of the first semester: the EE program of the nearby Fontys University of Applied Science accepts TU/e drop outs in their curriculum until the beginning of January, without any time losses involved. Besides, there is always a subset of students which the department considers a "risk group", i.e. students who may be successful but who need extra attention or specific individual care in order to succeed. Detecting this risk group in an early stage is essential for keeping these students from dropping out. It enables the department to direct its resources to the students who need it most. Current approach at EE department. To support students in making this decision, every enrolled student receives a study advice in December. This advice tells the student whether or not he or she is encouraged to proceed his study career at the faculty. It is based upon the grades and other results of the student so far and upon information obtained from 1st-semester-teachers and student-mentors, examined and interpreted by

41

Educational Data Mining 2009

the department's student counselor. The final semester examinations are not taken into account, because they are in January; postponing the advice until after the results are known would preclude students from switching to Fontys. The advices seem to be quite accurate in practice: students who are assessed as potentially successful are in general the same students that are successful after a year. Moreover, the students who are not encouraged to proceed their current study program, generally do not continue into the second year. The objectives. Despite the success, the assessment remains unsatisfactory because of its rather subjective character. Therefore, a more...
tracking img