wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Introduction to Machine Learning (Mid-term Exam)

Total questions: 40

Worksheet time: 40mins

Name
Class
Date
1.

What is the main goal of Machine Learning?

a)

To write programs manually for every possible case

b)

To enable computers to learn from data and make predictions

c)

To replace data with algorithms

d)

To store large amounts of information

2.

Which of the following is an example of supervised learning?

a)

Clustering customers by purchase history

b)

Predicting house prices based on features

c)

Detecting topics in a set of news articles

d)

Reducing image dimensions using PCA

3.

In supervised learning, the dataset includes:

a)

Only input variables

b)

Only output variables

c)

Both input and output variables

d)

Random unlabeled data

4.

Which of the following tasks best represents unsupervised learning?

a)

Predicting whether an email is spam or not

b)

Grouping customers based on their buying habits

c)

Predicting a student’s final grade

d)

Classifying handwritten digits

5.

What is “training data”?

a)

Data used to evaluate the final model

b)

Data used to build and adjust the model’s parameters

c)

Data that is randomly generated

d)

Data that contains only categorical variables

6.

Which of the following is NOT a type of Machine Learning paradigm?

a)

Supervised Learning

b)

Reinforcement Learning

c)

Controlled Learning

d)

Unsupervised Learning

7.

In reinforcement learning, the agent learns by:

a)

Being told the correct answers directly

b)

Maximizing cumulative reward through trial and error

c)

Randomly adjusting its parameters

d)

Following a fixed rule-based system

8.

Which of the following libraries is commonly used for Machine Learning in Python?

a)

NumPy

b)

Matplotlib

c)

Scikit-learn

d)

TensorDoc

9.

The process of dividing the dataset into training and testing sets is meant to:

a)

Reduce dataset size

b)

Evaluate model performance on unseen data

c)

Increase the training accuracy

d)

Avoid overfitting completely

10.

What does overfitting mean?

a)

Model fits the training data too well but fails to generalize

b)

Model performs poorly on both training and testing data

c)

Model is too simple for the given problem

d)

Model performs perfectly on new data

11.

Which metric is best suited for classification tasks?

a)

Mean Squared Error (MSE)

b)

Accuracy

c)

R-squared

d)

Mean Absolute Error (MAE)

12.

Which of the following evaluation techniques helps ensure robustness of model performance?

a)

Cross-validation

b)

Data augmentation

c)

Feature scaling

d)

Early stopping

13.

The confusion matrix is used for evaluating:

a)

Regression models

b)

Classification models

c)

Clustering models

d)

Dimensionality reduction

14.

Precision measures:

a)

The proportion of true positives among all predicted positives

b)

The proportion of true positives among all actual positives

c)

The overall accuracy of the model

d)

The false positive rate

15.

Recall is defined as:

a)

TP/(TP+FP)\mathrm{TP}/(\mathrm{TP}+\mathrm{FP})

b)

TP/(TP+FN)\mathrm{TP}/(\mathrm{TP}+\mathrm{FN})

c)

(TP+TN)/(TP+FP+FN+TN)(\mathrm{TP}+\mathrm{TN})/(\mathrm{TP}+\mathrm{FP}+\mathrm{FN}+\mathrm{TN})

d)

FP/(FP+TN)\mathrm{FP}/(\mathrm{FP}+\mathrm{TN})

16.

The ROC curve plots:

a)

Precision vs Recall

b)

True Positive Rate vs False Positive Rate

c)

Accuracy vs Loss

d)

Sensitivity vs Specificity

17.

In linear regression, the output variable must be:

a)

Categorical

b)

Continuous

c)

Binary

d)

Discrete

18.

The cost function commonly used in linear regression is:

a)

Cross-Entropy Loss

b)

Mean Squared Error (MSE)

c)

Hinge Loss

d)

Log Loss

19.

Which of the following statements about the regression line is true?

a)

It minimizes the distance between predicted and actual values

b)

It maximizes the variance of predictions

c)

It ignores the relationship between variables

d)

It always passes through the origin

20.

The slope (β1) in a simple linear regression represents:

a)

The average of Y

b)

The change in Y for a unit change in X

c)

The intercept

d)

The variance of X

21.

Logistic regression is mainly used for:

a)

Predicting continuous numeric outcomes

b)

Binary or categorical classification problems

c)

Clustering unlabeled data

d)

Detecting outliers

22.

The output of logistic regression before applying the sigmoid function is called:

a)

Logit

b)

Accuracy

c)

Loss

d)

Entropy

23.

The sigmoid function converts values into:

a)

0 or 1 directly

b)

Probability values between 0 and 1

c)

Negative values only

d)

Integer counts

24.

Which of the following activation functions is used in logistic regression?

a)

ReLU

b)

Sigmoid

c)

Tanh

d)

Softmax

25.

If the logistic regression model outputs 0.85, what does it mean?

a)

The model is 85% sure the sample belongs to class 1

b)

The model misclassified 85% of samples

c)

The learning rate is 0.85

d)

The accuracy is 0.85

26.

A Decision Tree splits data based on:

a)

Random numbers

b)

Feature values that minimize impurity

c)

Dataset size

d)

Alphabetical order

27.

Which metric is commonly used to measure impurity in a classification tree?

a)

Gini Index

b)

Mean Squared Error

c)

R² Score

d)

Gradient Loss

28.

What is a "leaf node" in a decision tree?

a)

A node that makes further splits

b)

The top-most node

c)

The final node that gives the output

d)

A node with maximum impurity

29.

Decision trees are prone to:

a)

Underfitting

b)

Overfitting

c)

High bias

d)

Slow learning

30.

Which technique helps reduce overfitting in Decision Trees?

a)

Increasing depth

b)

Pruning

c)

Using fewer features

d)

Removing labels

31.

In Random Forest, each tree is trained on:

a)

The same dataset

b)

Different random subsets of data and features

c)

Only the outliers

d)

A single feature

32.

The main idea behind ensemble methods is to:

a)

Combine multiple models to improve performance

b)

Use a single complex model

c)

Reduce dataset size

d)

Eliminate weak learners

33.

Bagging primarily helps to:

a)

Reduce variance

b)

Increase bias

c)

Increase training error

d)

Reduce number of features

34.

Boosting algorithms like AdaBoost and Gradient Boosting work by:

a)

Training models sequentially, focusing on previous errors

b)

Training all models independently

c)

Randomly selecting outputs

d)

Ignoring misclassified data

35.

A hyperparameter is:

a)

A parameter learned during training

b)

A parameter set before training begins

c)

Introduction to Machine

Learning

d)

Always related to data scaling

36.

Which of the following is a hyperparameter?

a)

Coefficient weight

( β\beta )

b)

Learning rate

c)

Predicted label

d)

Training accuracy

37.

The main goal of model optimization is to:

a)

Make model training slower

b)

Find the best set of parameters for minimal loss

c)

Reduce dataset size

d)

Simplify the algorithm

38.

Gradient Descent works by:

a)

Moving parameters in the direction of increasing loss

b)

Randomly updating weights

c)

Iteratively reducing the loss by adjusting weights

d)

Skipping local minima

39.

Early stopping is used to:

a)

Stop training when validation error starts increasing

b)

Stop training after a fixed number of epochs

c)

Stop when training accuracy is 100%

d)

Reset model weights

40.

Grid Search is a technique for:

a)

Randomly sampling hyperparameters

b)

Exhaustively testing all combinations of hyperparameter values

c)

Automatically generating datasets

d)

Visualizing model performance