Back to GalleryElastic Net
Elastic Net combines both L1 (Lasso) and L2 (Ridge) penalties. This is useful when there are multiple correlated features, as Lasso alone tends to select one feature randomly, whereas Elastic Net retains the group dynamics of correlated features.
Key Parameters & Visual Influence:
• Alpha (Penalty): Governs the overall magnitude of the regularization penalty.
• L1 Ratio: Adjusts the mix between L1 and L2 regularization. An L1 ratio of 1.0 represents pure Lasso, while 0.0 represents pure Ridge. Intermediate values allow balanced feature selection and coefficient shrinkage.
Dataset Studio
CSV format: Headers in first row, last column = label/target. All values must be numeric.
Example: x1, x2, label
Hyperparameters
Alpha (Penalty)Overall regularization penalty strength combining L1 and L2 penalties.1.0000
0.0001100
L1 RatioRatio of L1 Lasso vs L2 Ridge penalties (0 = pure Ridge, 1 = pure Lasso).0.50
01
Learning RateStep size taken during gradient descent optimization updates.0.0100
0.00011
Max EpochsMaximum training iterations over the dataset.300
102000
4 / 4 parameters active