Lasso Regressor

Lasso Regression adds L1 regularization to linear regression, which can shrink irrelevant feature coefficients to exactly zero. This means Lasso not only predicts values but also performs automatic feature selection.

It is the best choice when you suspect only a subset of your features is truly relevant.

Feature scaling is required Lasso is sensitive to feature scale. Standardization (e.g., StandardScaler) is usually necessary.

Regularization strength (alpha) tuning is critical

• If alpha is too large → underfitting.
• If alpha is too small → behaves like ordinary linear regression.

Use Cases

• Parametric design optimization
• Energy performance prediction: Predict building energy consumption based on geometric and spatial characteristics.
• Cost estimation models: Predict construction costs based on geometric measurements of building elements.
• Structural performance approximation: Predict structural performance based on geometric and spatial characteristics.