Regression

from SymbolicDSGE import regression
from SymbolicDSGE.regression import (
    RegressionKind,
    RegressionResult,
    RegressionStatus,
)

The regression subpackage contains the standard linear regression interfaces used directly by users and by Monte Carlo regression steps.

Public Modules:

Module	Description
`regression.ols`	Ordinary Least Squares regression and OLS inference diagnostics.
`regression.ridge`	Ridge regression and L2 grid search.
`regression.lasso`	Lasso regression and Lasso grid/path utilities.
`regression.elastic_net`	Elastic Net regression and grid search.
`regression.sr`	Symbolic-regression utilities.

RegressionKind Values:

Member	String	Dispatch Target
`OLS`	`"ols"`	`regression.ols.ols`
`RIDGE`	`"ridge"`	`regression.ridge.ridge`
`RIDGE_GS`	`"ridge_gs"`	`regression.ridge.ridge_gs`
`LASSO`	`"lasso"`	`regression.lasso.lasso`
`LASSO_GS`	`"lasso_gs"`	`regression.lasso.lasso_gs`
`ELASTIC_NET`	`"elastic_net"`	`regression.elastic_net.elastic_net`
`ELASTIC_NET_GS`	`"elastic_net_gs"`	`regression.elastic_net.elastic_net_gs`

RegressionStatus Values:

Member	Code	Description
`OK`	`0`	Regression solver completed normally.
`RANK_DEFICIENT`	`-1`	The primary linear solver detected rank deficiency and a fallback or failure path was used.
`NON_CONVERGENT`	`-2`	An iterative solver exhausted its iteration budget before satisfying tolerance.

Intercept Convention

Regression functions use intercept=True by default. When enabled, the returned design matrix contains an Intercept column as the first variable. Penalized methods do not penalize this intercept term.