KDnuggets Home » Polls » Data Mining Methodology (Aug 2007)

Data Mining Methodology


What main methodology are you using for data mining? [150 votes total]

CRISP-DM (63) 42%
My own (29) 19%
SEMMA (19) 13%
KDD Process (11) 7%
My organizations' (8) 5%
Domain-specific methodology (7) 5%
Other methodology, not domain-specific (6) 4%
None (7) 5%


Editor, Changes since 2004
Comparing the results to 2004 KDnuggets Poll on Data Mining Methodology, we see that exactly the same percentage (42%) chose CRISP-DM as the main methodology.
Among significant changes, percent who use their own methodology declined from 28% in 2004 to 19% in 2007, and percent who use SEMMA increased from 10% to 13%.

Simon Trippner, methodology
I use CRISP-DM because it is the most effective one. It is easy to apply and is very complex in scope. It is independent from any industry. For me CRISP-DM is the best way to follow a Data Mining process.

Omar Calzadilla, 3 methodology
I am using 3 methodology, CRISP-DM, Domain-specific methodology and My own, on medical databases.

KDnuggets Home » Polls » Data Mining Methodology (Aug 2007)