Motivation has a strong influence on students’ educational outcomes. This research aims to develop a framework to explore and predict computer science students’ motivation within the LA context based on Self-Determination Theory (SDT). The study has two main objectives: first, to explore students’ motivational behaviours using LA features, and second, to develop a model using Machine Learning (ML) techniques to predict those behaviours. Data were collected based on two research strategies which are case study and experiment. The case study targeted second and third-year computer science students. While the experiment targets all students who entered the department in 2018, 2019, and 2020. The purpose of the case study is to explore in deep the factors affecting CS students' motivation behaviours. On the other hand, the experiment aims to develop a model to explore and predict students' motivation types. Three machine learning (ML) models, including Multinominal Logistic Regression (MLR), Support Vector Machines (SVM), and XGBoost, were tested to classify motivational types into Intrinsic Motivation, Identified Regulation or Introjected Regulation. The performance of these models was evaluated using precision, recall and F1-scores. The XGBoost model, showed excellent accuracy in categorizing motivation behaviours. Therefore, the LA features resulting from the XGBOOST model were considered. For example, in the intrinsic motivation behaviours, five LA features were highlighted practical work, attendance, total duration, total activities and video. For the identified regulations behaviours, practical work, attendance, total duration, grade and assignment. For introjected regulations, practical work, total duration, grade, total clicks and video. After training the model, it was deployed in newly unseen data to predict students' motivation types and then determine the best LA features influencing each type of motivation. These results highlighted the possibility of effectively predicting student motivation within the LA context. The research highlights the importance of integrating advanced analytics in educational settings to foster student motivation.
| Date of Award | 1 Sept 2025 |
|---|
| Original language | English |
|---|
| Supervisor | Rupert Ward (Main Supervisor) & Richard Hill (Co-Supervisor) |
|---|