Rethinking big data: the role of artificial intelligence and machine learning

摘要

It seems that all current data management endeavours impinge on the capture and use of data - whether in science, engineering, industry or government - now consistently refer to "big data" if they refer at all to data. This use of the term "big data" is now common, but the breadth of possible interpretations is enormous. Included in this variety of interpretations are the identification of the attributes of velocity, variety, volume, veracity [Zikopoulos et al., 2005][Thanos et al., 2012], or the transition from OLTP-OLAP-RTAP in the evolution of data mining (e.g., [Heising, 2010]), or the technology-response definitions like MapReduce [Dean and Ghemawat, 2004] and HaDoop [Shafer et al., 2010], or even vendor-driven interpretations like the application development model of IBM's InfoSphere stream mining [Rea and Mamidipaka, 2010].