It is interesting to see that this topic which is around the academic world since 2004 is gaining popularity only now.
An interesting thing is that you can find some of these concepts implemented in apache spark (count approx) and into BlinkDB.
While, from a theoretical point of view many result are relate d to statistical bound like Chernoff bound and Chebyshev inequality.
During my phd I found this topic interesting and promising to answer to mine data with a verified error bound for precision/recall and elapsed time.
An interesting thing is that you can find some of these concepts implemented in apache spark (count approx) and into BlinkDB. While, from a theoretical point of view many result are relate d to statistical bound like Chernoff bound and Chebyshev inequality. During my phd I found this topic interesting and promising to answer to mine data with a verified error bound for precision/recall and elapsed time.
To go deeply into this topic I suggest the book Knowledge Discovery from Data Streams http://www.liaad.up.pt/area/jgama/DataStreamsCRC.pdf and other books written by Joao Gama.