Where to start

Applied Machine Learning

A short check to suggest a starting chapter. It's advice only: your answers aren't saved, nothing gets unlocked, and it doesn't count as evidence of a skill. It hasn't been tested on learners yet, and the questions are in English.

What's the gradient of f(w) = ‖Xw − y‖² with respect to w?

2 Xᵀ(Xw − y)

2 (Xw − y)

Xᵀ y

A fraud model reaches 99.5 % accuracy where 0.5 % of cases are fraud. What's your first reaction?

It's excellent and ready to deploy

It may be no better than always predicting "not fraud"

Accuracy can't exceed 99 %

What does increasing the penalty in ridge regression usually do?

Lowers training error

Removes the need for scaling features

Shrinks coefficients, trading some bias for less variance

What does it mean for a classifier to be well calibrated?

Among cases given probability 0.7, about 70 % are positive

Its accuracy is 70 % or more

Its threshold is 0.5

Why can a random split mislead when forecasting next month's sales?

Random splits use too little data

Training then includes information from after the test rows

Sales can't be predicted by models

Where should a scaler be fitted during cross-validation?

On all the data, before splitting

On the validation fold only

On the training fold only, inside each fold

Understand

Learn by doing. Understand why.

Learn

Explore the programmes Higher education High school Programming Data Science & AI Search

Understand

The method Create an account Log in

Help and privacy

Privacy Contact