Fast, Approximate Analysis of Big Data
The Challenge: Fast, Approximate Analysis of Big Data Computer Scientists have known about these types of queries for a long time, but not much attention was paid to the impact of these queries until the Internet exploded and Big Data reared its ugly head. It has been proved (and can be intuited, with some thought) that in order to compute these queries exactly, assuming nothing about the input stream and, for the quantiles case, without any restrictions on the number of quantiles requested, requires the query process to keep copies of every unique value encountered....