Image Dataset for Predicting Early Dropouts in DigitalLearning Platforms

This article presents a student click-stream database comprising of 120542 train images and 80362 test images where each directory contains two sub directories i.e. "Dropouts" and "NonDropouts" as two different classes.The original dataset was provided by KDD Cup Challenge 2015 in which the dataset was provided by chinese MOOC(Massive open online course) platform XuetangX. These samples have been acquired or captured through the clickstream activity/user activity on the platform. We transformed the KDD-Cup 2015 dataset into an image dataset. This transformation will enable the application of novel deep learning and computer vision techniques to develop more sustainable, accurate, and robust predictive models for identifying students at risk of dropping out and will enable MOOC platforms to design highly robust Early Warning Systems. Furthermore, this dataset will be made publicly available to the research community to advance interdisciplinary research at the intersection of education and computer vision.