Knowledge check

Check your knowledge

What is the difference between classical regression and classification models?

Regression models provide labels such as ‘cherry’/’banana’ but classification models calculate continuous numbers

Classification models and linear regression models are two names for the same thing

Classification models provide labels such as ‘cherry’/’banana’ but regression models calculate continuous numbers

How can we improve real-world model performance?

By adding features

Both adding and removing features can be useful, depending on the situation

By removing features

What is one reason why logistic regression uses log-loss rather than a more intuitive cost function?

Log-loss is stricter about model errors, even if they're close to correct

It's the only way to calculate error for categorical labels

Log-loss is more permissive to model error if that error is close to correct

You must answer all questions before checking your work.