Multiple choice › ML algorithms
ML algorithms
1. ML algorithms
Which problem is BEST solved with linear regression?
- Predicting the winner of a two-party election
- Deciding whether an email is spam
- Predicting the price of a house from its floor area
- Deciding whether a student is admitted to a course
Answer
C
Linear regression predicts a continuous numeric value. The other three are yes/no classifications (logistic regression). 2025 Q12 tested this.
2. ML algorithms
Logistic regression is most appropriate for:
- Predicting a binary outcome such as pass/fail
- Grouping unlabelled data
- Compressing images
- Predicting a continuous value such as temperature
Answer
A
Despite its name, logistic regression is a classification method that outputs a probability for a category (usually two).
3. ML algorithms
K-nearest neighbour (KNN) classifies a new item by:
- Looking at the labels of the K closest examples in the training data
- Randomly choosing a label
- Drawing a straight line through all the data
- Removing all labelled data
Answer
A
KNN is a supervised algorithm: the majority label of the K nearest neighbours is used. It is not a clustering algorithm (school notes say 'clustering such as KNN' which is incorrect).
4. ML algorithms
Polynomial regression is useful when:
- There are no numeric inputs
- The outcome must be yes or no
- The relationship between the variables is curved rather than a straight line
- The data is unlabelled
Answer
C
Polynomial regression fits a curve (x squared, x cubed...) to numeric data; it is still used to predict numeric values.