Long Live (Big Data-fied) Statistics!
Norman Matloff · 2013
Big Data is of intense interest today, in business, government, health care and so on. Though many of the problems are inherently statistical, the center of gravity of research in the field has moved to computer science. Indeed, some claim the topic to be outside the realm of statistics entirely. Yet it is shown here that Big Data makes statistics more important than ever. Statistical methods, and even more importantly, statistical concepts, are vital to effective use of Big Data. Equally, parallel processing methods are needed to make it all work. This paper takes a statistical approach to two major problems with Big Data: visualization and the Curse of Dimensionality.