Predictive Model for Classification of Student Learning Behaviour in E-Learning using English-Malay Code-Mixing Corpus at Institute of Higher Learning

Abstract

The current predictive model was developed using structured data by selecting irrelevant features due to measurement bias making it challenging for educators to effectively identify learning behaviour of the students. Therefore, this study focused on unstructured data which are code-mixed text that consisted of English and Malay from various E-learning platforms. In this study, An English-Malay code-mixing corpus was developed which associate with the predictive model to reflect the language patterns and communication styles often found among students in the e-learning platforms. To build the model, natural language processing techniques and algorithms was employed. After the preprocessing, the text data was annotated by psychology experts based on the four classification of student learning behaviour.  The result of the study includes the development of a robust predictive model that can effectively classify student learning behaviour in e-learning environments.