Predicting academic performance from behavioural and learning data

Abstract

The volume and quality of data, but also their relevance, are crucial when performing data analysis. In this paper, a study of the influence of different types of data is presented, particularly in the context of educational data obtained from Learning Management Systems (LMSs). These systems provide a large amount of data from the student activity but they usually do not describe the results of the learning process, i.e., they describe the behaviour but not the learning results. The starting hypothesis states that complementing behavioural data with other more relevant data (regarding learning outcomes) can lead to a better analysis of the learning process, that is, in particular it is possible to early predict the student final performance. A learning platform has been specially developed to collect data not just from the usage but also related to the way students learn and progress in training activities. Data of both types are used to build a progressive predictive system for helping in the learning process. This model is based on a classifier that uses the Support Vector Machine technique. As a result, the system obtains a weekly classification of each student as the probability of belonging to one of three classes: high, medium and low performance. The results show that, supplementing behavioural data with learning data allows us to obtain better predictions about the results of the students in a learning system. Moreover, it can be deduced that the use of heterogeneous data enriches the final performance of the prediction algorithms.

Student Engagement Predictions in an e-Learning System and Their Impact on Student Course Assessment Scores

“…Therefore, before applying the ML algorithms, we conducted a statistical analysis (Spearman correlation) to determine the significance between the dependent variable of the study (level of engagement) and the independent variables ( score on the assessment and the number of clicks on VLE activities, namely, dataplus , forumng , glossary , oucollaborate , oucontent , resource , subpage , homepage , and URL ). A spearman correlation is appropriate for both continuous and discrete features [ 78 ].…”

Analyzing and Predicting Students’ Performance by Means of Machine Learning: A Review

“…Bayesian Additive Regressive Trees (BART) was used to predict the final grade of students in the sixth week [30]. A model based on SVM weekly predicted the probability of each student belonging to one of these three types: high, medium or low performance [31]. Latent Dirichlet Allocation (LDA) predicted student grades according to how the students described their learning situations after each lesson [32].…”

“…It is possible to predict final students' performance beforehand thanks to behavioural data supplemented with other more relevant data (related to learning results). The system proposed in [31] obtained a weekly ranking of each student's probability of belonging to one of these three classification levels: high, medium or low performance. This performance could have something to do with non-cognitive characteristics which can have a significant impact on the students [9].…”

Use of Learning Analytics Interactive Dashboards in Serious Games

“…In order to verify the hypothesis that complementing behavioral data with other more relevant data (related to the learning results) can lead to a better analysis of the learning process, making it possible to predict the student's final performance in advance. The study showed satisfactory results and classified users with the probability of belonging to one of the three classes: high, medium and low performance (Villagrá-Arnedo et al, 2016). Studies in computational technologies also stand out, for addressing the development and application of assessment tools in serious games, such as the EngAGe tool, which is an integrated assessment tool for teachers and developers designed to separate the serious game from the system evaluation (Chaudy & Connolly, 2019).…”

“…The collection and analysis of data proves to be an extremely relevant tool among computational technologies, presenting behaviors of great significance for the learning process. In the study carried out in the Learning Management System, a learning platform was developed to collect data not only on the use of the system, but also related to the way students learn and progress training activities (Villagrá-Arnedo et al, 2016). In order to verify the hypothesis that complementing behavioral data with other more relevant data (related to the learning results) can lead to a better analysis of the learning process, making it possible to predict the student's final performance in advance.…”