Multimedia Social Big Data: Mining

Abstract

The rapid evolution and adoption of the SMAC (Social media, Mobile, Analytics and Cloud) technology paradigm, has generated massive volumes of human-centric, real-time, multimodal, heterogeneous data. Human-sourced information from social networks, process-mediated data from business systems and machine-generated data from Internet-of-Things are the three primary sources of big data which define the richness and scale of multimedia content available. With the proliferation of social networks (Twitter, Tumblr, Google+, Facebook, Instagram, Snapchat, YouTube, etc.), the user can post and share all kinds of multimedia content (text, image, audio, video) in the social setting using the Internet without much knowledge about the Web’s client-server architecture and network topology. This proffer novel opportunities and challenges to leverage high-diversity multimedia data in concurrence to the huge amount of social data. In recent years, multimedia analytics as a technology-based solution has attracted a lot of attention by both researchers and practitioners. The mining opportunities to analyze, model and discover knowledge from the social web applications/services are not restricted to the text-based big data, but extend to the partially unknown complex structures of image, audio and video. Interestingly, the big data is estimated to be 90% unstructured further, making it crucial to tap and analyze information using contemporary tools. The work presented is an extensive and organized overview of the multimedia social big data mining and applications. A comprehensive coverage of the taxonomy, types and techniques of Multimedia Social Big Data mining is put forward. A SWOT Analysis is done to understand the feasibility and scope of social multimedia content and big data analytics is also illustrated. Recent applications and suitable directions for future research have been identified which validate and endorse this correlation of multimedia to big data for mining social data.