Set
the the maximum tolerated squared Pearson correlation coefficient
r2 (with 0 ≤ r2 ≤ 1) between the prediction error on verification
data and the actual target values.
Notes: - r2=0 means that the prediction error and
the target field
values are uncorrelated, r2=1 means that the two values are
perfectly (positively or negatively) correlated. Therefore,
setting an upper threshold for r2 means setting an upper limit
for the acceptable systematic error of the model.
- Setting
a rSquared threshold is just a hint for the regression
algorithm. It cannot be guaranteed that the constructed model
really has a rSquared value below that threshold.
Parameters
- regSet
- A value of type DM_RegSettings
- regRSquared
- A value of type
DOUBLE.
Return
Value
The return Value is of type DM_RegSettings
- If regRSquared is NULL, regSet is returned. regSet did
not change.
- If regRSquared is not of type
DOUBLE or is invalid, an
error occurs.
- If none of the conditions above apply, the value regSet is
returned. The property is set to regRSquared.