Abstract
Clickbait is one of the form of false content, purposely designed to attract the user's attention and make them curious to follow the link and read, view, or listen to the attached content. The teaser aim behind this is to exploit the curiosity gap by giving information within the short statement. Still, the given statement is not sufficient enough to satisfy the curiosity without clicking through the linked content and lure the user to get into the respective page via playing with human psychology and degrades the user experience. To counter this problem, we develop a Clickbait Video Detector (CVD) scheme. The scheme leverages to learn three sets of latent features based on User Profiling, Video-Content, and Human Consensus, these are further used to retrieve cognitive evidence for the detection of clickbait videos on YouTube. The first step is to extract audio from the videos, which is further transformed to textual data, and later on, it is utilized for the extraction of video content-based features. Secondly, the comments are analyzed, and features are extracted based on human responses/reactions over the posted content. Lastly, user profile based features are extracted. Finally, all these features are fed into the classifier. The proposed method is tested on the publicly available fake video corpus [FVC], [FVC-2018] dataset, and a self-generated misleading video dataset [MVD]. The achieved result is compared with other state-of-the-art methods and demonstrates superior performance.