🤖 Data & AI

Machine learning foundations

Overfitting, train/test splits, metrics and regularization — the ideas every model question comes back to.

8questions

mediumdifficulty

+16max XP (1st try)

not rated yet

Question 1 of 8

A model with low training error and high test error is:

Question 2 of 8

Why must the test set be untouched during model selection?

Question 3 of 8

For a rare disease affecting 1 in 1000, a model that always predicts 'healthy' has:

Question 4 of 8

Precision answers:

Question 5 of 8

L2 regularization (ridge) reduces overfitting by:

Question 6 of 8

The bias–variance trade-off says a model that is too simple has:

Question 7 of 8

Cross-validation is used to:

Question 8 of 8

Feature scaling matters most for:

0/8 answered

Part of