Dakota Reference Manual  Version 6.15
Explore and Predict with Confidence
 All Pages
l2_penalty


Penalty value applied in regularized regression solver for function train approximations

Specification

Alias: none

Argument(s): REAL

Description

Regularized regression applies a penalty to aspects of the recovered solution (e.g., cardinality, higher-order oscillatory terms, et al.) to mitigate overfitting. This specification provides the scalar penalty parameter (its value is not currently adapted or cross-validated).

Default Behavior

If this penalty parameter is not specified, then we adopt the internal C3 default.

See Also

These keywords may also be of interest: