Visualizing the Hotspots and Emerging Trends of Multimedia Big Data through Scientometrics

摘要

Multimedia Big Data, known as the biggest big data, is becoming the forefront of big data research. However, a visualization research on the hotspots and trends of Multimedia Big Data through scientometric is still lacking. Based on the references from SCI-EXPANDED(SCIE), SSCI, CPCI-S, CPCI-SSHSI and arXiv databases in 2008-2017, the hotspots and emerging trends of Multimedia Big Data were identified for the first time by visualizing the co-cited references network, co-occurrence keywords network, burst references, burst keywords, Dual-Map Overlays network and Timeline networks with the information visualization software CiteSpaceV, Google Fusion Tables and Carrot(2). The results show that: (1)Multimedia Big Data research has spread across the globe, especially in the United States, China and some European countries; (2)"big data, web application, data mining, virtual screening, cloud service, structure-activity relationship, similarity search problems, concept modeling, etc. are the research hotspots; (3) the research focus evolved mainly from basic security problems and algorithm problems in the early, to technical problems, then to the applications and social impacts, and to mobile internet, cloud, data screening, payment security, etc. till now.(4) The emerging trends mainly include social influence modeling, mobile media cloud, video surveillance system, semantic relations, privacy, internet of thing, precision medicine, parallel massive clustering, etc.; (5) Multimedia Big Data research is developing toward interdisciplinary, of which mathematics and systems is a hot discipline and medicine and clinical is an emerging discipline; (6) the fusion development of multimedia big data with smart city, automotive industry, clothing industry and medical industry will be the trends of the times. The paper aims to promote the development of related theories on Multimedia Big Data and provide reference for researchers to identify relevant research directions.