40 Mining Your Own Business in Telecoms Using DB2 Intelligent Miner for Data
The last step is the final selection of features/variables for the mining run.
Variables could be superfluous by presenting the same or very similar
information as others, but increasing the run time.
Dependent or highly
correlated variables
could be found with statistical tests like bivariate statistics,
linear and polynomial regression. Dependent variables should be reduced by
selecting one variable for all others or by composing a new variable for all
correlated ones by factor or component analysis.
Not all variables remaining after the statistical check are nominated as input; only
variables with a clear interpretation and variables that make sense for the end ...