Context: This Research to Practice Full Paper presents a systematic review of methodologies that propose ways of reducing the dropout rate in Virtual Learning Environments (VLE). This generates large amounts of data about courses and students, whose analysis requires the use of computational analytical tools. Most educational institutions claim that the greatest issue in virtual learning courses is high student dropout rates. Goal: Our study aims to identify solutions that use Machine Learning (ML) techniques to reduce these high dropout rates. Method: We conducted a systematic review to identify, filter and classify primary studies. Results: The initial search of academic databases resulted in 199 papers, of which 13 papers were included in the final analysis. The review reports the historical evolution of the publications, the Machine Learning techniques used, the characteristics of data used, as well as identify solutions proposed to reduce dropout in distance learning. Conclusion: Our study provides an overview of the state of the art of solutions proposed to reduce dropout rates using ML techniques and may guide future studies and tool development. Keywords- dropout prediction; machine learning; distance education; online learning; a systematic review.
Distance education has surged with the gradual progress in information technology over the past decade and the Scholar Dropout phenomenon has been increasing, having repercussions in social, economic, and academic aspects, among others. These changes have created new challenges for different stakeholders in managing the learning process through a virtual platform. A Virtual Learning Environment (VLE) is a virtual classroom that allows teachers and students to communicate with each other online. Class information, learning materials, and assignments are provided via the Web. The amount of data collected through educational database technologies is increasing rapidly in volume and complexity, which allows for statistical analysis, data mining, and predictive actions.
In the last 10 years, Education Data Mining (EDM) has emerged as a new area concerned with the application of Data Mining (DM), Machine Learning (ML), and statistics of information generated from an educational setting , and the knowledge discovered may help improve teaching/learning processes. This study analyzed ML techniques in VLE. There are several studies that compare the performance of algorithms used in student performance prediction or dropout prediction systems, but this study looks for concrete solutions with a method that is designed or implemented using ML techniques. Thus, this paper aims to conduct a systematic review of the information obtained in the literature about solutions that reduce the high dropout rates, leading to the need for early identification of students who are likely to drop out and the possibility of providing teachers, tutors, and managers with strategic information to identify possible dropouts, which helps in decision making about adequate pedagogical intervention. This paper is organized as follows: Section II presents the background, Section III presents related works, Section IV details the research method, Section V presents data analysis, discusses the results, and presents an overview. Finally, section VI concludes the paper and presents further research and challenges of dropout prediction models using ML techniques.
- Hardware Requirement:
- OS – Windows 7, 8 and 10 (32 and 64 bit)
- RAM – 4GB
- Programming Language: