Towards Predicting Student’s Dropout in University Courses Using Different Machine Learning Techniques

Abstract

Early and precisely predicting the students’ dropout based on available educational data belongs to the widespread research topic of the learning analytics research field. Despite the amount of already realized research, the progress is not significant and persists on all educational data levels. Even though various features have already been researched, there is still an open question, which features can be considered appropriate for different machine learning classifiers applied to the typical scarce set of educational data at the e-learning course level. Therefore, the main goal of the research is to emphasize the importance of the data understanding, data gathering phase, stress the limitations of the available datasets of educational data, compare the performance of several machine learning classifiers, and show that also a limited set of features, which are available for teachers in the e-learning course, can predict student’s dropout with sufficient accuracy if the performance metrics are thoroughly considered. The data collected from four academic years were analyzed. The features selected in this study proved to be applicable in predicting course completers and non-completers. The prediction accuracy varied between 77 and 93% on unseen data from the next academic year. In addition to the frequently used performance metrics, the comparison of machine learning classifiers homogeneity was analyzed to overcome the impact of the limited size of the dataset on obtained high values of performance metrics. The results showed that several machine learning algorithms could be successfully applied to a scarce dataset of educational data. Simultaneously, classification performance metrics should be thoroughly considered before deciding to deploy the best performance classification model to predict potential dropout cases and design beneficial intervention mechanisms.

Is There a Real Need for the Preparatory Years in Higher Education? An Educational Data Analysis for College and Future Career Readiness

“…Through the intensive statistical analyses carried out by the researchers in this study, it was found that the CEAWR after the preparatory year is more convincing in assessing the performance of students, and therefore enrollment in colleges is more suitable for this performance. These findings align with other research findings that prove the usefulness of the first year on the prediction of students' success (Bruinsma and Jansen 2007;Van Der Zanden et al 2019;Beaulac and Rosenthal 2019;Culver and Bowman 2020;Kabathova and Drlik 2021).…”

“…The study also found that some factors, such as race, sex, and ACT score, may affect first-year seminars. In a recent study, Kabathova and Drlik (2021) discussed the use of learning analytics to explore the levels of student dropout from university studies using the available educational data. The study attempted to address appropriate features for machine learning classifiers used e-learning courses, which provide access to tests, assignments, exams, and projects.…”

The Effect of Resilience and Gender on the Persistence of Higher Education Students

“…It is often presumed by education policy and institutions that less persistent students are probably not sufficiently prepared. The review of the literature shows that many studies find a correlation between lower entry scores and failures in the first academic year, whereby a low level of persistence is developed due to poor academic performance (Kabathova and Drlik 2021;Reason 2009). In contrast, the first successful entry, favorable exam results, and good grades in the initial years all increase persistence.…”

An Early Intervention Technique for At-Risk Prediction of Higher Education Students in Cloud-based Virtual Learning Environment using Classification Algorithms during COVID-19

“…Most adopted algorithms are Logistic Regression, Decision Tree, Naive Bayes Classifier, Support Vector Machine, Random Forest, Neural Network. Results proved that Random Forest classifier obtained best accuracy with 93%, precision reached 86%, F1 score was 91% compared to other classifiers [26]. Researchers have collected real students' data with various information namely personal, economic, and academic records and evaluated by statistics values to find most effective one.…”