Supervised Learning
03 / 13
Linear Regression — Interview Questions
A curated bank of conceptual, mathematical, and case-study interview questions on Linear Regression.
How to use this page
Each question is tagged Easy / Medium / Hard. Try to answer before expanding the accordion. Questions are grouped by theme — Fundamentals, Regularization, Diagnostics, and Tricky cases.
1. Fundamentals & Math
2. Regularization (Ridge, Lasso, Elastic Net)
3. Assumptions & Diagnostics
4. Tricky / Case-Study Questions
Quick-fire one-liners
- Why intercept? So the model isn't forced through the origin.
- Bias-variance with λ? ↑λ → ↑bias, ↓variance.
- Cook's distance? Influence of a single point on all fitted values.
- Dummy variable trap? Drop one one-hot column to avoid perfect collinearity.
- Robust loss? Huber — quadratic near 0, linear in the tails.
# Quick sklearn cheat-sheet for interviews
from sklearn.linear_model import LinearRegression, Ridge, Lasso, ElasticNetCV
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
# OLS
ols = LinearRegression().fit(X, y)
# Ridge with CV — always scale first!
ridge = make_pipeline(StandardScaler(), Ridge(alpha=1.0)).fit(X, y)
# Lasso for feature selection
lasso = make_pipeline(StandardScaler(), Lasso(alpha=0.1)).fit(X, y)
# Elastic Net — handles correlated features + sparsity
enet = make_pipeline(StandardScaler(),
ElasticNetCV(l1_ratio=[.1, .5, .9], cv=5)).fit(X, y)