Machine Learning Viva Questions and

One-Line Answers (Q1 - Q70)

Machine Learning enables systems to learn from data and improve without being
explicitly programmed.

2. 2. Types of Machine Learning?

Supervised, Unsupervised, and Reinforcement Learning.

3. 3. What is Supervised Learning?

It uses labeled data to train models for prediction.

4. 4. What is Unsupervised Learning?

It finds hidden patterns in data without labels.

5. 5. What is Reinforcement Learning?

An agent learns to make decisions by rewarding correct actions.

6. 6. Example of Supervised Learning?

Spam detection using labeled email data.

7. 7. Example of Unsupervised Learning?

Customer segmentation using K-Means.

8. 8. What is Overfitting?

When a model memorizes training data and performs poorly on new data.

9. 9. What is Underfitting?

When a model is too simple and fails to capture data patterns.

10. 10. What is a Dataset?

A collection of data used to train or test ML models.

11. 11. What is a Feature?

An individual measurable property of the data.

The target or output value in supervised learning.

13. 13. What is Model Training?

The process where a model learns patterns from data.

14. 14. What is Model Testing?

Evaluating a trained model on unseen data.

15. 15. What is Linear Regression?

A model that predicts continuous values based on input features.

16. 16. What is Logistic Regression?

A model used for binary classification problems.

17. 17. What is KNN?

K-Nearest Neighbors classifies a data point based on its K closest neighbors.

18. 18. What is Decision Tree?

A tree-based model that splits data based on feature values.

19. 19. What is Random Forest?

An ensemble of decision trees used to improve accuracy.

20. 20. What is Naive Bayes?

A classification technique based on Bayes' theorem and feature independence.

21. 21. What is SVM?

Support Vector Machine finds the optimal boundary between classes.

22. 22. What is Clustering?

Grouping similar data points without labels.

23. 23. What is K-Means?

An algorithm that partitions data into K clusters based on distance.

24. 24. What is DBSCAN?

A clustering algorithm that groups data based on density.

Principal Component Analysis reduces data dimensions by retaining maximum variance.

26. 26. What is Feature Scaling?

Standardizing data to bring all features to a similar range.

27. 27. What is Normalization?

Scaling data between 0 and 1.

28. 28. What is Standardization?

Scaling data to have mean 0 and standard deviation 1.

29. 29. What is One-Hot Encoding?

Converting categorical variables into binary vectors.

30. 30. What is Train-Test Split?

Dividing data into training and testing sets for model evaluation.

31. 31. What is Cross-Validation?

It splits data multiple times to evaluate model performance reliably.

32. 32. Parametric vs Non-Parametric Models?

Parametric models have fixed parameters, non-parametric models adapt to data

33. 33. What is Data Preprocessing?

Cleaning and transforming raw data for ML models.

34. 34. What is Ensemble Learning?

Combining multiple models to improve prediction accuracy.

35. 35. What is Label Encoding?

Converting categorical values into numeric labels.

36. 36. PCA vs LDA?

PCA is unsupervised for variance; LDA is supervised for class separation.

37. 37. What is an Activation Function?

Adds non-linearity to neural networks for learning complex patterns.

Controls how much weights update during training.

39. 39. What is an Outlier?

A data point that significantly differs from others.

40. 40. Elbow Method in K-Means?

Used to find optimal K by identifying the point of diminishing returns.

41. 41. What is Bias and Variance?

Bias is underfitting error, variance is overfitting error; both must be balanced.

42. 42. What is Over/Under Sampling?

Techniques to handle class imbalance by altering data distribution.

43. 43. What is Overfitting Prevention?

Use regularization, cross-validation, or simpler models.

44. 44. What is Cross-Validation?

A method for assessing how the results of a model will generalize.

45. 45. Supervised vs Unsupervised?

Supervised uses labeled data; unsupervised doesn’t.

46. 46. Examples of Supervised Algorithms?

Linear Regression, Decision Tree, SVM, KNN.

47. 47. Examples of Unsupervised Algorithms?

K-Means, DBSCAN, PCA, Hierarchical Clustering.

48. 48. Hard vs Soft Clustering?

Hard assigns to one cluster; soft assigns probabilities.

49. 49. What is Feature Engineering?

Creating or modifying features to improve model performance.

50. 50. What are Precision, Recall, and F1-Score?

Precision: correct positives, Recall: found positives, F1: harmonic mean of both.
51. 51. What is Dropout?

Randomly disabling neurons to prevent overfitting in neural networks.

52. 52. Classification vs Regression?

Classification predicts categories; regression predicts numbers.

53. 53. What is SMOTE?

Synthetic Minority Over-sampling Technique balances data by generating synthetic

54. 54. What is a Confusion Matrix?

A table showing TP, FP, FN, TN for classification evaluation.

55. 55. What is a Learning Curve?

Graph showing model performance over time or data size.

56. 56. What is Reinforcement Learning?

Learning through feedback by maximizing rewards.

57. 57. What is the Curse of Dimensionality?

Too many features can make models less effective and complex.

58. 58. What is Regularization?

A method to reduce overfitting by adding penalty terms.

59. 59. What is a Hyperparameter?

Settings defined before training that guide model behavior.

60. 60. Bagging vs Boosting?

Bagging builds models independently; boosting builds them sequentially.

61. 61. What is a Decision Tree?

A model that splits data based on conditions into branches.

62. 62. What is Entropy in Decision Trees?

A measure of randomness or impurity in data.

63. 63. What is Gradient Boosting?

A boosting method where models fix previous errors via gradients.

Evaluates classification models using TP, TN, FP, FN.

65. 65. What is a Pipeline?

Automates end-to-end ML workflow from preprocessing to prediction.

66. 66. What is an Epoch?

One full pass through the training dataset during training.

67. 67. What is Tokenization in NLP?

Breaking text into smaller units like words or sentences.

68. 68. What is Model Generalization?

Model's ability to perform well on unseen data.

69. 69. What is Bias-Variance Tradeoff?

Balancing underfitting and overfitting for optimal performance.

70. 70. Batch Size vs Epoch?

Batch size is the number of samples per training step; epoch is one full pass.