Data Mining is compression, and compression is undecidable, therefore ...
Date:
 Christos Faloutsos
Christos Faloutsos was the winner of
2010 SIGKDD Innovation Award
In his very lively and fun talk entitled
Large Graph Mining
Data Mining for fun (and profit)
(after thanking many colleagues, nominators, teachers, students, family,
funding agencies, ...)
Christos presented his thesis:
1. Data mining = compression
(Christos Faloutsos, Vasileios Megalooikonomou: On data
mining, compression, and Kolmogorov complexity. Data Min.
Knowl. Discov. 15(1): 3-20 (2007) )
2. Compression is undecidable
Theorem: for an arbitrary string x, computing
its Kolmogorov complexity K(x) is undecidable
3. Data miners have job security (:-)
since there will always be better data
mining tools/models/patterns to be discovered
He also showed many interesting examples of models, such as the graph of Metabolic rate vs Mass which shows very unusual patterns found in real data.
Here is
PDF of Prof. Faloutsos talk
|