Course Project

About this page

This page is a brief summary of the course project for this course. The project was done in three phases:

• Descriptive Analytics: The learners were expected to come up with insightful and meaningful visualisations for the OULAD dataset.

• Prediction of course outcome: In this phase, the learners were required to early predict the course outcome of the learners (Pass, Fail, Withdrawn, Distinction)

• Open-Ended Question: The learners were expected to identify some research gap from literature on this dataset, come up with their own research questions and use suitable methods to answer them appropriately.

About the dataset

The dataset used for this project is the publicly available Open University Learning Analytics Dataset (OULAD). It contains data about courses, students and their interactions with Virtual Learning Environment (VLE) for seven selected courses. You can learn more about the dataset on: https://analyse.kmi.open.ac.uk/open_dataset

DeSCRIPTIVE ANALYtics

PREDICTIVE ANALYTICS

Summary of the work done

• The following is a list of features (not exhaustive) that the students have explored for the prediction of course outcome:

Course module ID, course presentation ID, region, highest education, number of previous attempts, the sum of clicks in the VLE, Mean number of clicks in the VLE, date submitted, forum access, glossary access, homepage access, resource access, studied credits, IMD band, Age band, disability, Tutor Marked Assessment, Computer Marked Assessment.

• The students tried out a number of classifiers ranging from Decision Trees, Random Forest, Naive Bayes, Logistic Regression, AdaBoost Classifier to Neural Networks.

• Students were asked to perform 10-fold cross-validation and predict the course outcome into 4 categories- pass, fail, withdrawn, distinction.

• The best-reported accuracy on the given task was 95%.