Linear Regression
Predicts a continuous output as a weighted sum of features, fit by minimizing squared errors.
- Standard fit is OLS: closed-form, unbiased under classical assumptions.
- Coefficients are directly interpretable.
- Foundation for GLMs, regularization, and most regression metrics.
When to use: Strong baseline whenever the target is continuous; reach for it before anything fancier.
Example: House prices: price = 50000 + 120 · sqft + 8000 · bedrooms − 500 · age. Each coefficient is interpretable in dollars per unit.