A survey of Machine Learning Approaches and Techniques for Student Dropout Prediction

Tracking #: 535-1515


Responsible editor: 

Richard Mann

Submission Type: 

Survey Paper

Abstract: 

School dropout is absenteeism from school for no good reason for a continuous number of days. Ad- dressing this challenge requires a thorough understanding of the underlying issues and effective planning for interventions. Over the years machine learning has gained much attention in addressing the problem of students dropout. This is because machine learning techniques can effectively facilitate determination of at-risk students and timely planning for interventions. To this end, several machine learning algorithms have been proposed in literature. This paper presents an overview of machine learning in education with focus on approaches and techniques for student-dropout prediction. Furthermore, the paper discusses the state of student dropout in developing countries and several performance metrics used by researchers to evaluate machine learning techniques in the context of education with experimental examples. Finally, the paper highlights challenges and future research directions.

Manuscript: 

Tags: 

  • Reviewed

Data repository URLs: 

Date of Submission: 

Sunday, June 10, 2018

Date of Decision: 

Wednesday, July 4, 2018

Decision: 

Undecided

Solicited Reviews:


2 Comments

I think this paper is trying

I think this paper is trying to cover a bit too much. Perhaps it would be better to either dive into a couple of methods and how they have been used to predict student dropout or review the results of papers utilizing machine learning methods. Right now, the paper really glosses over the algorithms and relative strengths/weaknesses of each approach. It also glosses over the results and insight gained from studies that have used machine learning to understand drop rates. I'd suggest either focusing on the algorithms in more depth (perhaps comparing methods on a single dataset to compare efficacy of each) or structuring as a meta-analysis of prior studies that have used machine learning methods.

Meta-Review by Editor

Your manuscript has been reviewed by four reviewers, who had mixed views on the paper. 

In general the reviewers thought that the study covered an important area where there has been little previous study, and where research is needed. Therefore this study could potentially be an important contribution to the literature.

However, the referees also flagged up several areas of concern, which would need to be fully addressed in a revised version of the manuscript. I give a summary of the major themes below, but a revised manuscript should respond to each individual reviewer’s comments

  • The paper needs careful proofing to remove typographic and grammatical errors. Several reviewers found that these impeded their understanding of the work
  • The value of the new data analysis presented in the paper needs to be made clearer in terms of what it adds to the literature survey. As the paper has been submitted as ‘survey paper’, new research should illuminate points made in the survey of the existing literature. Some concerns were also raised about the details of your machine learning analysis and experimental design, which should be clarified and justified.

Given the potential importance of this work, I would welcome submission of a revised manuscript that addresses all the reviewers' concerns.

Richard Mann (http://orcid.org/0000-0003-0701-1274)