wayground logo

Free Printable Worksheets

NEW

Font size

S
M
L
XL
Worksheets

Machine learning quiz

Total questions: 17

Worksheet time: 9mins

Name
Class
Date
1.

Q.1) Which of the following best describes the goal of linear regression?

a)

A. Classify data into categories

b)

B. Predict a continuous numerical value

c)

C. Cluster data points into groups

d)

D. Reduce the dimensionality of data

2.

Q2 Decision trees split nodes based on which criterion to maximize separation of classes?

a)

A. Euclidean distance

b)

B. Gini impurity / Information gain

c)

C. Gradient descent

d)

D. K-distance

3.

Q3 Which of the following is not an assumption of classical linear regression?

a)

A. Linearity between predictors and response

b)

B. Independence of errors

c)

C. Homoscedasticity (constant variance of errors)

d)

D. The predictors must all be integers

4.

Q4 Which metric is not typically used to evaluate regression models?

a)

A. Mean Absolute Error (MAE)

b)

B. R-squared

c)

C. Adjusted R-squared

d)

D. F1-score

5.

Q5 In a decision tree for classification, common impurity measures for splitting are:

a)

A. Gini impurity and entropy

b)

B. Euclidean distance and Manhattan distance

c)

C. Mean squared error and R-squared

d)

D. Gradient descent and backpropagation

6.

Q6 Random Forest improves accuracy mainly by:

a)

A. Using one deep tree on the whole dataset

b)

B. Combining many trees trained on bootstrapped samples

c)

C. Choosing medoids instead of centroids

d)

D. Applying L1 regularization to each tree

7.

Q7 Gradient Boosting differs from Bagging because it:

a)

A. Builds trees in paralle

b)

B. Builds trees sequentially, each correcting the previous one

c)

C. Uses k-nearest neighbors for splits

d)

D. Requires distance metrics

8.

Q7 Which statement about decision trees is true?

a)

A. They can only handle numerical data

b)

B. They require feature scaling

c)

C. They can naturally handle both categorical and numerical features

d)

D. They cannot be used for regression

9.

Q. Pruning a decision tree is primarily done to:

a)

A. Increase training accuracy

b)

B. Reduce overfitting and improve generalization

c)

C. Make the tree deeper

d)

D. Convert it into a random forest

10.

Q. Which of the following is not a type of machine learning?

a)

A. Supervised

b)

B. Unsupervised

c)

C. Reinforcement

d)

D. Compilation

11.

Q. K-Nearest Neighbors (KNN) is a:

a)

A. Parametric model

b)

B. Non-parametric model

c)

C. Ensemble model

d)

D. Dimensionality-reduction method

12.

Q. Which of these algorithms is not distance-based?

a)

A. K-Nearest Neighbors (KNN)

b)

B. K-Means

c)

C. Naïve Bayes

d)

D. DBSCAN

13.

Q. The most commonly used distance metric in Euclidean space is:

a)

A. Cosine similarity

b)

B. Manhattan distance

c)

C. Euclidean distance

d)

D. Hamming distance

14.

Q. In KNN, the parameter k refers to:

a)

A. Number of clusters

b)

B. Number of neighbors considered when predicting a label

c)

C. Number of features in the dataset

d)

D. Number of decision trees

15.

Q. Feature scaling (e.g., standardization) is critical for KNN because:

a)
It increases the number of features
b)
It ensures all features are binary
c)
It eliminates the need for cross-validation
d)
A. Distances are affected by different feature scales
16.

Q. Which voting scheme is common in KNN classification?

a)
A. Majority vote among the k neighbors
b)
Average of k neighbor votes
c)
Sum of neighbor labels
d)
Random selection of a neighbor
17.

Q. K-Means clustering minimizes:

a)
A. Mean absolute error
b)
B. Sum of squared Euclidean distances to cluster centroids (Answer)
c)
D. Total variance
d)
C. Logarithmic loss