LASSO + ELASTIC NET
Follow the regularization path.
Watch features enter a model as the penalty relaxes. The Rust solver runs right here, in your browser.
LOCAL FILE
Tell us what to fit.
All other response labels map to 0; the complete column must contain exactly two classes.
Choose predictors
Preview of the first rows. Every selected value is checked during import. Missing values and categorical predictors need preprocessing.
THE SOLUTION PATH
From sparse to expressive.
A whole family of models, in one fit.
Load a dataset to trace its coefficients.
Then move along λ to explore the trade-off.
Paths show up to 24 features with the largest maximum standardized coefficients. The table and exports include every fitted feature.
CROSS-VALIDATION
THE SELECTED MODEL
Coefficients on your data’s scale.
Sorted by absolute coefficient. These magnitudes depend on feature units; the plot above uses standardized coefficients.
| Feature | Coefficient | Standardized |
|---|
These are training fits, not an estimate of predictive performance. Turn on cross-validation for one.