VetaQL: Video Metadata Querying System Powered with Video Recommendation

Increasingly video sharing and OTT platforms aim to hyper-personalize their content for the users. This paper proposes a novel approach to dynamically recommend videos or similar events in other videos to users based on the detailed context at a particular time in a video. This concept opens up a new dimension for video recommendation. We ascertain the exact details of the video by modeling the video metadata into a generalized event model semantically, then storing this semantically linked event modeled metadata efficiently to recommend videos/events to users and demonstrate semantic video querying through the proposed VetaQL language.