Detection and Prevention of Copywriting Multimedia Contents in Youtube Using Multi-Keyword Ranking Algorithm

摘要

YouTube, an American video sharing website has become the most preferred site for watching videos online, with millions of content creators. YouTube is often said to be the second largest search engine in the world after Google itself, where people spend hours watching videos, generating billions of views. Popularity dynamics of YouTube videos highly depends on the meta-tags, number of view counts and the social dynamics which indicates the interaction between the content creators (channels) and YouTube users. Social networks like videos spreads and sensitivity of YouTube meta-level features provide an important impact on the popularity of videos. Our dataset includes all the videos from the YouTube site and are imported into a centralized database of our system using EmbetiSource: In the context of video analysis, each video is characterized into four attributes: Title, Category, Description, Embedded links. The non-frequent search words used by the users are converted to its equivalent signature using the Digital Signature Algorithm (DSA). We propose a Multi -Keyword Ranking (MKR) algorithm which scales the popular or originally created video using attributes such as view count, number of subscribers, title, description and the comments and ratings generated by users computed using sentimental analysis. The outcome of this proposed work would be an optimized search of the original video that was actually created by the channel.