Back to Subject

Unit 3: SUPERVISED LEARNING: CLASSIFICATION - Practice Quiz

INT234 — Predictive Analytics 50 Questions
0 Correct 0 Wrong 50 Left
0/50

1 Which of the following algorithms is categorized as a 'Lazy Learner'?

A. K-Nearest Neighbors
B. Support Vector Machines
C. Naïve Bayes
D. Decision Trees

2 In K-Nearest Neighbors, what is the likely effect of choosing a very small value for 'k' (e.g., k=1)?

A. High variance and overfitting
B. High bias and low variance
C. The model becomes too simple
D. The model ignores local patterns

3 Which distance metric is most commonly used in k-NN for continuous numerical variables?

A. Euclidean distance
B. Jaccard similarity
C. Hamming distance
D. Cosine similarity

4 Why is feature scaling (normalization/standardization) important in k-NN?

A. To reduce the size of the dataset
B. To convert categorical data to numerical
C. To increase the value of k
D. To prevent features with larger scales from dominating the distance calculation

5 The 'Naïve' in Naïve Bayes refers to which fundamental assumption?

A. All features are conditionally independent given the class
B. The prior probabilities are equal
C. The algorithm is simple to implement
D. All features are dependent on each other

6 Naïve Bayes is based on which mathematical theorem?

A. Pythagorean Theorem
B. Central Limit Theorem
C. Taylor's Theorem
D. Bayes' Theorem

7 What is the purpose of Laplace Smoothing in Naïve Bayes?

A. To normalize the data
B. To handle missing values
C. To handle the problem of zero probability for unseen features
D. To reduce the dimensionality

8 Which strategy is primarily used to build Decision Trees?

A. Backpropagation
B. Lazy learning
C. Divide and Conquer
D. Gradient Descent

9 In a Decision Tree, what does a leaf node represent?

A. A feature to split on
B. The root of the tree
C. A class label or decision
D. The entropy value

10 Which metric is commonly used to measure impurity in Decision Trees?

A. R-squared
B. Gini Index
C. Correlation Coefficient
D. Euclidean distance

11 The process of removing branches from a decision tree to prevent overfitting is called:

A. Scaling
B. Boosting
C. Regularization
D. Pruning

12 Which concept represents the expected reduction in entropy caused by partitioning the examples according to an attribute?

A. Information Gain
B. Log Loss
C. Gini Impurity
D. Maximum Margin

13 In rule-based classification, what does 'coverage' refer to?

A. The number of features used
B. The accuracy of the rule
C. The complexity of the rule
D. The number of instances that satisfy the rule's condition

14 The OneR (One Rule) algorithm generates rules based on:

A. The single most informative attribute
B. A random attribute
C. The nearest neighbors
D. All attributes simultaneously

15 What is the primary objective of a Support Vector Machine (SVM)?

A. Minimize the number of features
B. Maximize the posterior probability
C. Create the deepest possible decision tree
D. Find a hyperplane that maximizes the margin between classes

16 The data points that lie closest to the decision boundary in an SVM are known as:

A. Support Vectors
B. Outliers
C. Centroids
D. Noise

17 What technique does SVM use to handle non-linearly separable data?

A. Kernel Trick
B. Pruning
C. Bagging
D. Smoothing

18 In SVM, what is the role of the 'C' hyperparameter?

A. It determines the number of kernels
B. It calculates the Euclidean distance
C. It controls the trade-off between maximizing the margin and minimizing classification errors
D. It sets the depth of the tree

19 In a Confusion Matrix, what does 'False Positive' (Type I Error) represent?

A. Incorrectly predicting the positive class when it is actually negative
B. Correctly predicting the positive class
C. Incorrectly predicting the negative class when it is actually positive
D. Correctly predicting the negative class

20 Which formula correctly calculates Accuracy?

A. (TP + TN) / (TP + TN + FP + FN)
B. TP / (TP + FP)
C. TP / (TP + FN)
D. 2 (Precision Recall) / (Precision + Recall)

21 Accuracy is often a misleading metric when:

A. The dataset is imbalanced
B. The dataset is perfectly balanced
C. The dataset is small
D. The model is a decision tree

22 Which metric represents the ratio of correctly predicted positive observations to the total predicted positives?

A. Recall
B. Accuracy
C. Specificity
D. Precision

23 Recall is also known as:

A. F1 Score
B. Sensitivity
C. Specificity
D. Precision

24 The F1 Score is the harmonic mean of which two metrics?

A. Precision and Recall
B. True Positive Rate and False Positive Rate
C. Accuracy and Error Rate
D. Sensitivity and Specificity

25 Which metric would be most important for a spam detection system where it is acceptable to miss some spam, but critical not to delete legitimate emails (high cost of False Positive)?

A. Sensitivity
B. Log Loss
C. Recall
D. Precision

26 Which metric would be most important for cancer detection where missing a positive case is dangerous (high cost of False Negative)?

A. Accuracy
B. Precision
C. Recall
D. Specificity

27 What does AUC stand for in the context of model evaluation?

A. Average Unit Cost
B. Accuracy Under Classification
C. Area Under the Curve
D. Algorithm User Context

28 The ROC curve plots which two metrics against each other?

A. Accuracy vs Loss
B. Sensitivity vs Specificity
C. True Positive Rate vs False Positive Rate
D. Precision vs Recall

29 An AUC score of 0.5 indicates:

A. A model with zero error
B. A model with high precision
C. A model that predicts randomly
D. A perfect model

30 Logarithmic Loss (Log Loss) penalizes a classifier based on:

A. The depth of the tree
B. The number of misclassifications only
C. The confidence of the predicted probabilities
D. The number of support vectors

31 What is the ideal value for Logarithmic Loss?

A. 0
B. 1
C. 100
D. 0.5

32 Which of the following is a disadvantage of Decision Trees?

A. Cannot handle categorical data
B. Difficult to interpret
C. Prone to overfitting if not pruned
D. Requires feature scaling

33 Which algorithm is generally considered a 'Black Box' model due to low interpretability?

A. Rules (RIPPER)
B. Decision Trees
C. Linear Regression
D. Support Vector Machines (with RBF kernel)

34 In Naïve Bayes, what is the 'Posterior Probability'?

A. The probability of the evidence given the class
B. The initial probability of the class
C. The probability of the class given the evidence
D. The probability of the evidence regardless of class

35 Which classification algorithm is parametric?

A. K-Nearest Neighbors
B. Decision Trees
C. None of the above
D. Naïve Bayes

36 What is the 'Hinge Loss' function associated with?

A. K-Means
B. Logistic Regression
C. Support Vector Machines
D. Decision Trees

37 What happens to the computational cost of k-NN during the prediction phase as the dataset size grows?

A. It increases significantly
B. It decreases
C. It remains constant
D. It becomes zero

38 Entropy in Information Theory is a measure of:

A. Accuracy
B. Margin width
C. Disorder or Uncertainty
D. Distance

39 RIPPER (Repeated Incremental Pruning to Produce Error Reduction) is an algorithm used for:

A. Regression
B. Dimensionality Reduction
C. Rule Induction
D. Clustering

40 Which of the following describes a 'False Negative' (Type II Error)?

A. Predicting Negative when actually Negative
B. Predicting Positive when actually Negative
C. Predicting Positive when actually Positive
D. Predicting Negative when actually Positive

41 If Precision = 1.0 and Recall = 1.0, what is the F1 Score?

A. 0
B. 1.0
C. 2.0
D. 0.5

42 Which evaluation metric calculates the proportion of actual negatives that are correctly identified?

A. Recall
B. Specificity
C. Sensitivity
D. Precision

43 In a decision tree, if a node contains only samples from a single class, its entropy is:

A. Infinite
B. 1
C. 0
D. 0.5

44 Which kernel is the default for non-linear SVMs in many libraries?

A. Polynomial
B. Sigmoid
C. Linear
D. Radial Basis Function (RBF)

45 Generative models like Naïve Bayes model:

A. The error gradients
B. The boundary between classes directly
C. The distance between points
D. The distribution of individual classes (Joint probability)

46 Recursive Partitioning is a technique synonymous with:

A. Calculating Bayes probabilities
B. Building Decision Trees
C. Calculating k-NN distances
D. Optimizing SVM margins

47 When interpreting a Confusion Matrix for a multi-class problem (e.g., 3 classes), the matrix dimensions are:

A. 1x3
B. 3x1
C. 3x3
D. 2x2

48 Which algorithm is most sensitive to outliers?

A. Decision Trees
B. Naïve Bayes
C. Rules
D. K-Nearest Neighbors

49 What is the relationship between Error Rate and Accuracy?

A. Error Rate = Accuracy
B. Error Rate = 1 + Accuracy
C. Error Rate = 1 - Accuracy
D. Error Rate = Accuracy / 2

50 The 'Zero Frequency' problem in Naïve Bayes is solved using:

A. Pruning
B. Kernel Trick
C. Laplace Smoothing
D. Feature Scaling
Previous MCQ
Unit 2
Next MCQ
Unit 4