Early Student-at-Risk Detection by Current Learning Performance and Learning Behavior Indicators

Abstract

The article is focused on the problem of early prediction of students’ learning failures with the purpose of their possible prevention by timely introducing supportive measures. We propose an approach to designing a predictive model for an academic course or module taught in a blended learning format. We introduce certain requirements to predictive models concerning their applicability to the educational process such as interpretability, actionability, and adaptability to a course design. We test three types of classifiers meeting these requirements and choose the one that provides best performance starting from the early stages of the semester, and therefore provides various opportunities to timely support at-risk students. Our empirical studies confirm that the proposed approach is promising for the development of an early warning system in a higher education institution. Such systems can positively influence student retention rates and enhance learning and teaching experience for a long term.

Examining the factors affecting students' science success with Bayesian networks

“…BN, which is widely used in a variety of fields, has been used in a small number of studies in the field of education (Almond et al, 2015;Culbertson, 2016;Reichenberg, 2018). However, BN is more advantageous than other methods with its ability to model students in the field of education (Levy, 2016;Lytvynenko et al, 2019;Sinharay, 2006) and to evaluate the model quickly (Kenekayoro, 2018;Kustitskaya et al, 2020;Millán et al, 2013;Nguyen & Do, 2009). Essential improvements in the education system are vital to enhance students' success.…”

Big Data in Education: Students at Risk as a Case Study

“…Their approach achieved 89% accuracy and proposed personalized learning paths to boost academic success. Authors in [23] developed an early detection method for identifying at-risk students based on learning performance and behavior indicators, achieving 79% accuracy. Authors in [24] employed educational data mining and ML techniques to predict academic performance, with Decision Tree (DT) and ANN models achieving 82.38% and 84.57% accuracy, respectively.…”

“…Improved prediction of student failure [23] Early student-at-risk detection k-Nearest Neighbors (kNN) classifier and Linear Discriminant Analysis (LDA) Good detection performance [24] Predicting the academic performance of students RF, nearest neighbor, Support Vector Machines (SMVs), LR, Naïve Bayes, kNN Classification accuracy of 70-75% [25] Predicting MOOC performance and improving performance Bayesian Network Improved MOOC performance prediction [26] Create an online academic performance prediction system SVM, RF, kNN, ANN, and LR Performance was not the target of the paper. They used mean square error to measure the performance [27] Improved feature selection to predict student performance Enhanced Binary Genetic Algorithm (EBGA) Improved prediction accuracy [28] Predict student academic performance using supervised learning kNN, SVM, DT, NB, ANNs…”

"I feel like a bother": COVID-Era Experiences of College Students on Academic Probation and Financial Aid Warning Status

“…Institutions should continue exploring new communication methods to reach students from at-risk populations and encourage them to seek and integrate support services. Kustitskaya et al (2022) recommended that institutions engage in various assessment tools for student performance, emphasizing the first half of the semester. Perhaps instructors or case managers working for higher education institutions could explore a student readiness measure or assessment that evaluates students' willingness to seek and accept support services, identifying students who do not communicate or avoid institutional support.…”