Module linear 

Linear models: OLS, Ridge, Logistic, Lasso, and ElasticNet.

Regularization naming convention

scry-learn uses alpha as the regularization strength parameter across all linear models — matching scikit-learn’s Ridge, Lasso, and ElasticNet:

Model	Parameter	Meaning
LinearRegression	alpha	L2 penalty strength (0 = OLS)
Ridge	alpha	L2 penalty strength (constructor arg)
LassoRegression	alpha	L1 penalty strength
ElasticNet	alpha	Total penalty strength
LogisticRegression	alpha	Penalty strength (type set by Penalty)

sklearn migration note

scikit-learn’s LogisticRegression and SVC use C = 1/alpha (inverse regularization strength). When porting sklearn code, convert via alpha = 1.0 / C. All other sklearn linear models (Ridge, Lasso, ElasticNet) already use alpha, so those translate directly.

Structs

ElasticNet

ElasticNet regression (mixed L1 + L2 regularization).

LassoRegression

Lasso regression (L1-regularized linear regression).

LinearRegression

Linear regression model.

LogisticRegression

Logistic regression for binary/multiclass classification.

Ridge

Ridge regression — LinearRegression with L2 regularization.

Enums

Penalty

Regularization penalty for logistic regression.

Solver

Solver algorithm for logistic regression.