Evaluation Metrics for Linear Regression

What evaluation metrics do you use for linear regression, and what are the assumptions of linear regression?

Asked in the Technical round stage. Asked as one question in the only technical round.

Answer the question as you would in an interview. Explain which metrics you would select for different objectives, distinguish prediction quality from inferential validity, and describe how you would diagnose violations of the regression assumptions. Where relevant, connect model evaluation to experiment analysis without inventing product or business context.

1. Define appropriate regression metrics and their limitations.
2. State the core assumptions of ordinary least squares regression.
3. Describe diagnostics or remedies for assumption violations.
4. Explain how metric choice differs for prediction versus causal or experimental analysis.