Model Evaluation Metrics — ML Technical Interview Questions & Answers

Accuracy, precision, recall, F1 and RMSE.

Model Evaluation Metrics

  • Simple meaning: A metric is the mark you give the model. Classification: accuracy, precision, recall, F1. Regression: MAE, MSE, RMSE, R². Pick the metric that matches the job, not the one that looks biggest.
  • Picture: 95 students pass, 5 fail. A dummy that always says “pass” is 95% accurate and useless for catching fails. Use recall/F1 on fail, or a confusion matrix.
  • Regression picture: off by 2 marks vs off by 20. RMSE punishes big mistakes more than MAE.
  • Trap: accuracy on an imbalanced sheet. Trap: quoting R² without saying what 1.0 and 0.0 mean.

Technical Interview Tip: Class job → precision/recall/F1. Number job → MAE/RMSE. Accuracy is not always enough.

Code Examples for Model Evaluation Metrics

Model Evaluation Metrics — working example

y_true = [1, 0, 1, 1, 0, 1]
y_pred = [1, 0, 0, 1, 0, 1]
correct = sum(t == p for t, p in zip(y_true, y_pred))
accuracy = correct / len(y_true)
print(correct, accuracy)
# Accuracy alone misleads on imbalanced classes — check precision/recall too.

Model Evaluation Metrics — accuracy = correct / total. For imbalanced data, also report precision/recall/F1.

How to Explain Model Evaluation Metrics to an Interviewer

Accuracy trap?

Imbalanced classes Always-guess majority looks good

Name two regression metrics.

MAE RMSE R²