Visualizing student learning progression pattern using neural network and hierarchical clustering analysis

Motivating students to learn and continuous improvement of course materials are necessary and vital for the benefits of the students who participate in online courses. It is crucial to provide guidance on how to effectively use the course materials. This study aims to develop a method to identify progression patterns among students who successfully completed the course. This can help students to navigate through the course and instructors to understand more on how to organize the course materials. Traditional mining techniques such as prefix span can mine student’s platform usage patterns, yet, they do not provide a complete learning progression sequence of how course materials are related to one another. To collectively identify the learning progression pattern of successful candidates, this study proposes a descriptive model using a neural network with hierarchical clustering method. We fragment the learning path of successors into triplets and use a neural network to learn their inter-relationship in an unsupervised manner. The inter-relationship can then be visualized through fixing pre-test as the start point of the sequence to post-test as the endpoint based on the hierarchical clustering results. We have applied this technique to 3 classes of students in a statistics course. The learning sequence of successful learners revealed by the proposed method are similar across cohorts. Copyright © 2018 The Organization Committee of ISET 2018.