Features
From: Gregory Piatetsky-Shapiro
Date: 16 Feb 2005
Subject: Popular data mining techniques
The previous KDnuggets Poll asked:
What data mining/analytic techniques you use frequently?
The top 4 techniques were
- Decision Trees/Rules - 14%
- Clustering - 13%
- Regression - 11%
- Statistics - 10%
Here are full results for
2005 Poll on Data Mining Techniques.
Methods that gained popularity include
- Genetic algorithms
- SVM
- Bayesian
- Association rules
while methods that lost popularity include
- Neural networks
- Statistics
- Nearest neighbor
- Decision Trees
However, overall the results are surprisingly similar to 2003 Poll on Popular Data Mining Techniques. The top 2 methods are the same - Decision/trees and Clustering.
If Statistics and Regression are combined in 2005 poll, then their popularity is very similar to combined percentage of Statistics and Linear regression in 2003 poll.
Have data mining methods reached a plateau ?
|