Learner behavior prediction in a learning management system

Abstract

Learning Management Systems (LMS) lack automated intelligent components that analyze data and classify learners in terms of their respective characteristics. Manual methods involving administering questionnaires related to a specific learning style model and cognitive psychometric tests have been used to identify such behavior. The problem with such methods is that a learner can give inaccurate information. The manual method is also time-consuming and prone to errors. Although literature reports complex models predicting learning styles, only a few have used machine learning methods such as an artificial neural network (ANN). The primary objective of this study was to design, develop, and evaluate a model based on machine learning for predicting learner behavior from LMS log records. Approximately 200,000 log records of 311 students who had accessed e-Learning courses for a 15-week semester were extracted from LMS to create a dataset. Machine learning concepts were identified from the log records. The dataset was split into training and testing sets. A model using the artificial neural network algorithm was designed and implemented using an r-studio programming language. The model was trained to predict learner behavior and classify each student. The prediction success rate of 0.63, 0.67, 0.64, 0.65, 0.26, 0.64 accuracy, precision, recall, f-score, kappa, and Area Under the Curve (AUC) respectively were recorded. This demonstrates that the model after full validation can be relied on to identify learner behavior.