Machine Learning Quiz

This Machine Learning Quiz contains 30 multiple-choice questions covering core concepts, learning types, common algorithms, model evaluation, data preparation, and introductory deep learning. Each question has one correct answer, followed by an answer key with brief explanations.

How to Use This Machine Learning Quiz

• Answer all 30 machine learning MCQs before checking the answer key.
• Choose the single best answer for each question.
• Use the explanations to review concepts you missed.
• Give yourself one point for every correct answer.

Machine Learning Fundamentals MCQs

1. What is the main goal of machine learning?

1. To write a separate rule for every possible input
2. To enable systems to learn patterns from data and make predictions or decisions
3. To replace all database systems
4. To increase computer storage capacity

2. Which statement best describes a feature in machine learning?

1. A measurable input variable used by a model
2. The final prediction produced by a model
3. A hardware component used for training
4. A programming error in the dataset

3. In supervised learning, what is included with each training example?

1. Only an input value
2. An input and its corresponding target or label
3. Only a cluster identifier generated by the model
4. A random reward signal

4. Which problem is a classification task?

1. Predicting tomorrow’s temperature in degrees
2. Estimating the price of a house
3. Determining whether an email is spam or not spam
4. Forecasting monthly electricity consumption

5. Which problem is a regression task?

1. Predicting a continuous house price
2. Assigning a document to one of four topics
3. Identifying whether a transaction is fraudulent
4. Grouping customers without predefined labels

6. What is a training dataset primarily used for?

1. To estimate model parameters from examples
2. To store only the final predictions
3. To replace the test dataset
4. To remove every unusual observation automatically

7. What does a machine learning model learn during training?

1. Relationships in the training data that help map inputs to outputs
2. The physical location of each data file
3. Only the names of the input columns
4. A fixed answer that is used for every input

8. Which term refers to the correct output associated with a supervised training example?

1. Label
2. Cluster
3. Epoch
4. Hyperplane

9. What is inference in machine learning?

1. Using a trained model to produce predictions for input data
2. Deleting the training dataset after every epoch
3. Manually assigning labels to all future examples
4. Converting a regression problem into a database query

10. Which of the following is a hyperparameter?

1. A learning rate selected before or during model tuning
2. A weight learned directly from training data
3. A predicted class label
4. An observed target value

Supervised, Unsupervised, and Reinforcement Learning Questions

11. Which algorithm is commonly used for both binary classification and probability estimation?

1. Logistic regression
2. K-means clustering
3. Principal component analysis
4. Apriori association rule mining

12. What does a decision tree use to divide the data?

1. Feature-based split conditions
2. Only random labels
3. A single fixed threshold for every feature
4. Database joins

13. What is the main idea behind a random forest?

1. Combine predictions from multiple decision trees
2. Use a single tree with no branching
3. Group data into clusters using only Euclidean distance
4. Reduce every dataset to one feature

14. Which algorithm assigns observations to one of K groups without using class labels?

1. K-means clustering
2. Linear regression
3. Logistic regression
4. Naive Bayes classification

15. What is the objective of principal component analysis?

1. Represent data with fewer uncorrelated components while retaining as much variance as possible
2. Assign a class label to every observation
3. Increase the number of duplicate features
4. Generate labels for reinforcement learning

16. In reinforcement learning, what guides an agent’s behavior?

1. Rewards and penalties received from interactions with an environment
2. A complete set of correct labels for every state
3. Only a test dataset
4. A fixed cluster assignment

17. Which learning approach is most suitable for discovering customer segments when no segment labels exist?

1. Unsupervised learning
2. Supervised regression
3. Supervised classification
4. Rule-based sorting only

18. Which method predicts a target by examining nearby training examples?

1. K-nearest neighbors
2. Principal component analysis
3. K-means clustering
4. Gradient calculation only

19. What is an ensemble method?

1. A method that combines multiple models to produce a final prediction
2. A method that always uses one feature
3. A process for removing labels from data
4. A database normalization technique

20. Which algorithm tries to find a separating boundary with the largest margin between classes?

1. Support vector machine
2. K-means clustering
3. Linear regression
4. Principal component analysis

Model Evaluation and Data Preparation MCQs

21. Why is a test dataset kept separate from the training dataset?

1. To estimate how the trained model performs on unseen data
2. To make the training dataset larger
3. To guarantee perfect accuracy
4. To select labels after deployment

22. What is overfitting?

1. Performing very well on training data but poorly on unseen data
2. Performing equally on every possible dataset
3. Using too few input columns in all cases
4. Training without any model parameters

23. What is underfitting?

1. Failing to capture important patterns even in the training data
2. Memorizing every training example exactly
3. Using a separate validation set
4. Reducing the learning rate during tuning

24. Which metric is calculated as correct predictions divided by all predictions?

1. Accuracy
2. Precision
3. Recall
4. Mean squared error

25. Which metric answers: “Of all examples predicted as positive, how many were actually positive?”

1. Precision
2. Recall
3. Specificity
4. Mean absolute error

26. Which metric answers: “Of all actual positive examples, how many did the model identify?”

1. Recall
2. Precision
3. R-squared
4. Silhouette score

27. What is the purpose of cross-validation?

1. To evaluate a model across multiple train-validation splits
2. To ensure the model sees the test labels during training
3. To remove every missing value automatically
4. To convert all features into text

28. Why are numerical features often standardized?

1. To place features on comparable scales for algorithms sensitive to magnitude
2. To convert every numerical feature into a label
3. To guarantee that all algorithms become linear
4. To eliminate the need for evaluation

29. What does one-hot encoding commonly do?

1. Represent categories with separate binary indicator columns
2. Replace all numerical values with their mean
3. Remove every categorical feature
4. Sort records by their target value

30. In a neural network, what is an epoch?

1. One complete pass through the training dataset
2. One input feature
3. One predicted class
4. One hidden neuron only

Machine Learning Quiz Answer Key with Explanations

Question	Correct answer	Explanation
1	2	Machine learning uses data to learn patterns that support predictions or decisions instead of requiring a manually written rule for every case.
2	1	A feature is an input variable supplied to a model, such as age, distance, or transaction amount.
3	2	Supervised learning uses labeled examples containing both input data and a known target.
4	3	Spam detection predicts one of two categories, so it is a classification problem.
5	1	House price is a continuous numerical value, which makes this a regression task.
6	1	The training set is used to fit the model’s learnable parameters.
7	1	Training adjusts model parameters so the model can capture useful relationships between inputs and outputs.
8	1	A label or target is the known output associated with a supervised example.
9	1	Inference is the use of a trained model to generate a prediction for new or existing input data.
10	1	The learning rate is configured rather than directly learned as a model weight.
11	1	Logistic regression estimates class probabilities and is widely used for binary classification.
12	1	A decision tree repeatedly applies feature-based conditions to divide observations into branches.
13	1	A random forest aggregates predictions from many decision trees, usually reducing the instability of a single tree.
14	1	K-means is an unsupervised algorithm that partitions observations into K clusters.
15	1	PCA projects data onto orthogonal components ordered by the variance they explain.
16	1	A reinforcement learning agent learns from reward signals received after actions in an environment.
17	1	When no segment labels exist, clustering and other unsupervised methods can be used to discover structure.
18	1	K-nearest neighbors predicts using the labels or values of nearby training examples.
19	1	Ensemble methods combine several models, such as trees in a random forest or sequential learners in boosting.
20	1	A support vector machine seeks a separating hyperplane with a large margin between classes.
21	1	A separate test set provides a less biased estimate of performance on data not used to fit the model.
22	1	Overfitting occurs when a model learns training-specific detail that does not generalize well.
23	1	An underfit model is too limited, insufficiently trained, or otherwise unable to represent important patterns.
24	1	Accuracy is the fraction of all predictions that are correct.
25	1	Precision measures the proportion of predicted positives that are true positives.
26	1	Recall measures the proportion of actual positives identified by the model.
27	1	Cross-validation repeats training and validation across different partitions to estimate model performance more reliably.
28	1	Standardization can prevent large-scale features from dominating distance-based or gradient-based algorithms.
29	1	One-hot encoding creates a binary column for each represented category.
30	1	An epoch is one full pass through the training data during model training.

Machine Learning Quiz Scoring Guide

Score	Interpretation	Recommended review
26–30	Strong understanding of introductory machine learning concepts	Move to algorithm tuning, feature engineering, and applied projects.
20–25	Good foundation with a few gaps	Review the explanations for missed questions and compare related algorithms.
13–19	Developing understanding	Revisit supervised learning, unsupervised learning, and evaluation metrics.
0–12	More foundation work is needed	Study basic terminology, task types, datasets, and model validation before retaking the quiz.

Machine Learning Quiz Concepts to Review

• Classification: predicts a category, such as spam or not spam.
• Regression: predicts a continuous numerical value, such as price or demand.
• Clustering: groups similar observations without predefined class labels.
• Overfitting: occurs when training performance is strong but generalization is weak.
• Precision and recall: describe different aspects of classification performance and should be selected according to the cost of false positives and false negatives.
• Feature preprocessing: may include scaling, categorical encoding, missing-value treatment, and leakage prevention.

Machine Learning Quiz FAQs

What topics are covered in this machine learning quiz?

The quiz covers machine learning terminology, supervised and unsupervised learning, reinforcement learning, classification, regression, clustering, common algorithms, model evaluation, preprocessing, overfitting, and basic neural-network training.

Are these machine learning MCQs suitable for beginners?

Yes. The questions focus on introductory concepts and include brief explanations. Learners should still review practical examples because a multiple-choice quiz cannot test every implementation detail.

What is the difference between classification and regression in machine learning?

Classification predicts a discrete category, while regression predicts a continuous numerical value. Spam detection is a classification example, and house-price prediction is a regression example.

Why can accuracy be misleading for an imbalanced dataset?

A model can achieve high accuracy by repeatedly predicting the majority class while failing to identify the minority class. Precision, recall, F1 score, a confusion matrix, and class-specific results can provide a more useful evaluation.

How should I prepare for a machine learning assessment?

Review core definitions, understand when common algorithms are used, practise interpreting evaluation metrics, and work through small datasets. Focus on reasoning about model behavior rather than memorizing isolated answers.

Machine Learning Quiz Editorial QA Checklist

• Verify that each machine learning MCQ has exactly one clearly correct answer.
• Check that classification, regression, clustering, and reinforcement learning are described without mixing their objectives.
• Confirm that precision and recall definitions use the correct positive-prediction and actual-positive denominators.
• Ensure that test data is not described as data used to fit or tune the final model.
• Review algorithm descriptions for scope: logistic regression for classification, K-means for clustering, and PCA for dimensionality reduction.
• Retest the answer-key numbering after adding, removing, or reordering any question.