1
Q
What is the basic idea of classification_
A
- prediction/target y can only take on a small range of possible values
- like is a mail spam, yes or no
- linear regression is bad for this
2
Q
Why is linear regression less useful for classification?
A
- because outliers can easily influencing the function in a bad way
- because we need to use a threshold / decision boundary to declare when something belongs to what class
2
Q
What algorithm is widely used for binary classification problems?
A
- logistic regression
3
Q
A
Decks in class (11)
# Cards
