Leveraging Video Clickstream data to predict the user's Video Interest in a SPOC

In this work, we propose a model for predicting the user's interest in a video while enrolling in a Small Private Online Course (SPOC) by using the Bayes' theorem. We suggest tracking the following user's interactions with videos at the click level: Play, Pause, Move Back, Move Forward, Replay and Download. We consider that observable behaviors like viewing a video for a long time, replaying or uploading a video could be implicit indicators of the user's interest on videos. Our approach is a user-based method that leverages clickstream data gathered from the user's interactions and falls into the implicit approaches that infer the interest of users without asking them additional actions.