Week 2

Question 1

What is generalization?

Answer 1

How well our model can predict unseen data.

Question 2

What is the difference between regression and classification?

Answer 2

Classification is about predicting the class label from a pre-defined set of outcomes. This can be a binary outcome or a multiway outcome.

Regression is about predicting a continuous number. For example, predicting someone’s income from their age, education and job.

Question 3

What is overfitting?

Answer 3

Overfitting is when the model doesn’t generalize new data. It models the training set too well and can’t accurately predict new data.

Question 4

What is underfitting?

Answer 4

This is when the model is too simple. It can’t accurately predict the training data.

Question 5

What do you get if a model is too complex?

Answer 5

Overfitting

Question 6

What do you get if a model is too simple?

Answer 6

Underfitting.