Lecture 3::::::

1. What is the primary focus of Lecture 3 in machine

learning?
A) Model deployment
B) Advanced neural networks
C) Feature selection techniques
D) Supervised learning principles
Answer: D) Supervised learning principles

2. Which of the following is a key aspect of supervised

learning?
A) Uses unlabeled data
B) Clusters data into groups
C) Requires labeled input data
D) Applies unsupervised algorithms
Answer: C) Requires labeled input data

3. In supervised learning, what is a ‘label’?

A) A data point used for model training
B) The expected output for an input
C) A numerical value for features
D) An unstructured data type
Answer: B) The expected output for an input

4. Which algorithm is commonly used for regression tasks?

A) K-means clustering
B) Decision tree
C) Logistic regression
D) Linear regression
Answer: D) Linear regression

5. What is the role of ‘features’ in a machine learning model?

A) The raw input values
B) Parameters used in optimization
C) Variables influencing the prediction
D) Output values of the model
Answer: C) Variables influencing the prediction

6. What is a characteristic of a classification task?

A) Predicts continuous values
B) Sorts data into discrete categories
C) Clusters data without labels
D) Applies regression analysis
Answer: B) Sorts data into discrete categories
7. Which metric is commonly used to evaluate classification
accuracy?
A) Mean Absolute Error
B) Mean Squared Error
C) Precision
D) Recall
Answer: C) Precision

8. What does ‘overfitting’ mean in the context of machine

learning?
A) The model performs equally on all data
B) The model generalizes well
C) The model fits too closely to the training data
D) The model underestimates variability
Answer: C) The model fits too closely to the training data

9. Which technique is used to reduce overfitting?

A) Increasing data size
B) Adding more features
C) Removing the regularization parameter
D) Reducing training iterations
Answer: A) Increasing data size

10. In machine learning, what is ‘cross-validation’?

A) Splitting data into train and test sets
B) Testing different models on the same data
C) A method to improve data quality
D) A technique for assessing model stability
Answer: D) A technique for assessing model stability

11. What type of model would you use for a binary

classification task?
A) K-means clustering
B) Logistic regression
C) Linear regression
D) Principal Component Analysis
Answer: B) Logistic regression

12. What is ‘feature scaling’ used for in model training?

A) To reduce model complexity
B) To scale data to a common range
C) To add more features to the data
D) To remove outliers from the data
Answer: B) To scale data to a common range
13. Which algorithm is best suited for handling non-linear
relationships?
A) Linear regression
B) Logistic regression
C) Decision tree
D) Naive Bayes
Answer: C) Decision tree

14. In classification, what does ‘recall’ measure?

A) The number of false negatives
B) The proportion of true positives correctly identified
C) The model’s overall accuracy
D) The number of correct predictions
Answer: B) The proportion of true positives correctly identified

15. Which method helps select the best subset of features

for a model?
A) Regularization
B) Feature scaling
C) Feature extraction
D) Feature selection
Answer: D) Feature selection

16. What is ‘dimensionality reduction’ used for in machine

learning?
A) To increase the number of features
B) To simplify the data by reducing features
C) To split data into different clusters
D) To create additional training data
Answer: B) To simplify the data by reducing features

17. What is a hyperparameter?

A) A value that the model learns during training
B) A parameter used only in evaluation
C) A model setting adjusted before training
D) A variable specific to test data
Answer: C) A model setting adjusted before training

18. In supervised learning, what is the purpose of a ‘test

set’?
A) To train the model
B) To adjust hyperparameters
C) To evaluate model performance
D) To create a new dataset
Answer: C) To evaluate model performance
19. Which metric is most appropriate for regression tasks?
A) Accuracy
B) Recall
C) Mean Absolute Error (MAE)
D) F1 Score
Answer: C) Mean Absolute Error (MAE)

20. What does ‘bias’ refer to in machine learning models?

A) Random errors in predictions
B) A model’s preference for specific data points
C) The tendency of a model to underfit
D) Noise in the dataset
Answer: C) The tendency of a model to underfit

I’ll continue with the next set of 20 questions for a total of 40.

Here is the second set of 20 questions to complete the 40-question

set, ensuring comprehensive coverage of key concepts in machine
learning:

21. What does ‘variance’ in machine learning refer to?

A) The model’s capacity to adapt to new data
B) The variability of predictions for a given data point
C) The average difference between observed and predicted values
D) The total number of features in the data
Answer: B) The variability of predictions for a given data point

22. Which method helps avoid overfitting by averaging

multiple models?
A) Ensemble learning
B) Dimensionality reduction
C) Regularization
D) Cross-validation
Answer: A) Ensemble learning

23. In a decision tree, what is a ‘leaf node’?

A) The root of the tree
B) The point where branches split
C) The final output or prediction point
D) A node used to test features
Answer: C) The final output or prediction point

24. What is the purpose of regularization in machine

learning?
A) To increase the complexity of the model
B) To scale features uniformly
C) To reduce the risk of overfitting
D) To enhance feature selection
Answer: C) To reduce the risk of overfitting

25. In which scenario would you apply logistic regression?

A) When predicting continuous values
B) For unsupervised clustering
C) For binary classification problems
D) To analyze principal components
Answer: C) For binary classification problems

26. What is the main advantage of using a neural network?

A) Easy to interpret model structure
B) Fast to train
C) Good at handling complex non-linear relationships
D) Low computational requirements
Answer: C) Good at handling complex non-linear relationships

27. Which technique divides data into groups based on

similarities?
A) Classification
B) Regression
C) Clustering
D) Feature extraction
Answer: C) Clustering

28. In supervised learning, what is a validation set used for?

A) To adjust model parameters during training
B) To evaluate the model after training
C) To create labels for the test set
D) To pre-process the input data
Answer: A) To adjust model parameters during training

29. What is ‘model interpretability’?

A) The ease of applying the model to new data
B) The transparency of the model’s decision-making process
C) The speed of the model’s predictions
D) The total accuracy of the model
Answer: B) The transparency of the model’s decision-making
process

30. Which of the following is an ensemble method?

A) Decision tree
B) Neural network
C) Random forest
D) Logistic regression
Answer: C) Random forest

31. In a confusion matrix, what does a ‘true positive’

represent?
A) A negative instance predicted as positive
B) A positive instance correctly predicted as positive
C) A negative instance correctly predicted as negative
D) A positive instance predicted as negative
Answer: B) A positive instance correctly predicted as positive

32. Which algorithm is primarily used for anomaly detection?

A) K-means clustering
B) Support vector machine
C) Linear regression
D) PCA (Principal Component Analysis)
Answer: B) Support vector machine

33. What is a limitation of using neural networks?

A) Limited to linear data
B) High interpretability
C) Requires large datasets and high computational power
D) Easily overcomes overfitting issues
Answer: C) Requires large datasets and high computational power

34. Which concept describes the trade-off between bias and

variance?
A) Model interpretability
B) Hyperparameter tuning
C) Model complexity
D) Generalization
Answer: D) Generalization

35. What is ‘early stopping’ used for?

A) To halt training when performance decreases
B) To prevent the model from starting training
C) To increase computational efficiency
D) To improve test accuracy after training
Answer: A) To halt training when performance decreases

36. Which learning type uses feedback to optimize actions?

A) Supervised learning
B) Unsupervised learning
C) Reinforcement learning
D) Transfer learning
Answer: C) Reinforcement learning
37. In machine learning, what is ‘feature extraction’?
A) Selecting the most important features
B) Transforming raw data into new attributes
C) Grouping data by similarities
D) Balancing the dataset
Answer: B) Transforming raw data into new attributes

38. What is the main goal of transfer learning?

A) To use pre-trained models on new tasks
B) To optimize existing algorithms
C) To create a validation dataset
D) To label data for supervised learning
Answer: A) To use pre-trained models on new tasks

39. Which technique is commonly used to handle imbalanced

datasets?
A) Cross-validation
B) Oversampling
C) Feature extraction
D) Dimensionality reduction
Answer: B) Oversampling

40. In neural networks, what is the purpose of an ‘activation

function’?
A) To train the model faster
B) To map input to output directly
C) To introduce non-linear properties into the network
D) To reduce the dataset size
Answer: C) To introduce non-linear properties into the network