Back to Subject

Unit 4: UNSUPERVISED LEARNING: CLUSTERING AND PATTERN DETECTION - Practice Quiz

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

1 Which of the following best defines Unsupervised Learning?

A. Learning where the data has no predefined labels or target variables
B. Learning where the model predicts a continuous target variable
C. Learning where the model is rewarded or punished based on actions
D. Learning where the model predicts a categorical target variable

2 In K-Means clustering, what does 'K' represent?

A. The number of iterations
B. The distance metric used
C. The number of features in the dataset
D. The number of clusters the algorithm forms

3 What is the primary objective function that K-Means minimizes?

A. Inter-cluster distance
B. The Silhouette coefficient
C. Within-Cluster Sum of Squares (WCSS)
D. The number of outliers

4 Which of the following is the first step in the standard K-Means algorithm?

A. Calculate the centroid of all points
B. Calculate the total variance
C. Assign points to the nearest cluster
D. Randomly initialize K centroids

5 The 'Random Initialization Trap' in K-Means refers to:

A. Different initial centroid positions leading to different, suboptimal results
B. The inability to handle categorical data
C. The algorithm selecting the wrong number of K
D. The algorithm running indefinitely without convergence

6 Which technique is commonly used to mitigate the Random Initialization Trap?

A. K-Means++
B. Principal Component Analysis
C. Gradient Descent
D. Agglomerative Clustering

7 What is the 'Elbow Method' used for in K-Means clustering?

A. Visualizing high-dimensional data
B. Stopping the algorithm early
C. Determining the optimal number of clusters
D. Calculating the distance between centroids

8 In an Elbow Method plot, what variable is typically on the Y-axis?

A. Computation Time
B. WCSS (Inertia)
C. Accuracy
D. Number of Clusters (K)

9 When does the K-Means algorithm stop iterating?

A. When every point is in its own cluster
B. When the WCSS becomes zero
C. When the number of clusters equals the number of data points
D. When the centroids no longer move significantly between iterations

10 Which of the following is a limitation of K-Means clustering?

A. It is computationally expensive for small datasets
B. It requires the number of clusters to be specified in advance
C. It always finds the global optimum
D. It can only handle binary data

11 Agglomerative Hierarchical Clustering is best described as a approach.

A. Density-based
B. Bottom-up
C. Centroid-based
D. Top-down

12 Divisive Hierarchical Clustering is best described as a approach.

A. Randomized
B. Bottom-up
C. Top-down
D. Grid-based

13 What diagram is commonly used to visualize Hierarchical Clustering?

A. Scatter plot
B. Dendrogram
C. Histogram
D. Box plot

14 In a dendrogram, the vertical axis typically represents:

A. The time taken to cluster
B. The frequency of data points
C. The Euclidean distance or dissimilarity between clusters
D. The number of clusters

15 How do you determine the optimal number of clusters using a dendrogram?

A. It is impossible to determine K from a dendrogram
B. Choose the height where the first merge occurs
C. Count the number of leaves at the bottom
D. Cut the dendrogram at the point with the longest vertical distance without crossing horizontal lines

16 Which linkage method defines the distance between two clusters as the shortest distance between any single point in one cluster and any single point in the other?

A. Complete Linkage
B. Average Linkage
C. Single Linkage
D. Ward's Method

17 Which linkage method defines the distance between two clusters as the maximum distance between any point in the first cluster and any point in the second?

A. Single Linkage
B. Average Linkage
C. Centroid Linkage
D. Complete Linkage

18 Average Linkage calculates the distance between clusters by:

A. Using the distance between the centroids of the clusters
B. Using the minimum distance between points
C. Taking the average of all pairwise distances between points in the two clusters
D. Taking the median distance of all points

19 Centroid Linkage measures the distance between clusters based on:

A. The sum of squared errors
B. The furthest points in the clusters
C. The distance between the geometric centers (means) of the clusters
D. The closest points in the clusters

20 Which linkage method is most notorious for producing the 'chaining' effect (long, stringy clusters)?

A. Average Linkage
B. Single Linkage
C. Complete Linkage
D. Ward's Method

21 What is a primary advantage of Hierarchical Clustering over K-Means?

A. It works better with high-dimensional data
B. It is computationally faster on large datasets
C. It always uses Manhattan distance
D. It does not require assuming the number of clusters (K) beforehand

22 Market Basket Analysis is a specific application of which technique?

A. Clustering
B. Linear Regression
C. Association Rule Learning
D. Decision Trees

23 In the rule {Bread} -> {Butter}, {Bread} is the:

A. Support
B. Consequent
C. Lift
D. Antecedent

24 What does the metric 'Support' measure in Association Rules?

A. The frequency with which an itemset appears in the dataset
B. The correlation between items
C. The ratio of the rule's confidence to the expected confidence
D. The reliability of the rule

25 How is 'Confidence' for the rule A -> B calculated?

A. Support(A & B) / Support(A)
B. Support(A) / Support(B)
C. Support(A & B) / Support(B)
D. Support(B) / Support(A)

26 What does a 'Lift' value greater than 1 indicate?

A. The items are independent of each other
B. The rule is invalid
C. The items are substitutes (negatively correlated)
D. The presence of the antecedent increases the likelihood of the consequent

27 If the Lift of a rule A -> B is exactly 1, what does this imply?

A. A and B are independent
B. A and B are perfectly correlated
C. The confidence is 100%
D. A and B are never bought together

28 Which algorithm is most commonly associated with mining frequent itemsets for Association Rules?

A. Random Forest
B. K-Nearest Neighbors
C. Naive Bayes
D. Apriori

29 The Apriori algorithm uses the 'Downward Closure Property'. What does this property state?

A. If an itemset is infrequent, its subsets must be frequent
B. Support always equals Confidence
C. All supersets of a frequent itemset must be frequent
D. All subsets of a frequent itemset must be frequent

30 Which of the following is NOT a step in the K-Means algorithm?

A. Assignment of points to the nearest centroid
B. Update of centroids to the mean of assigned points
C. Calculation of distance matrix for all pairs of points
D. Initialization of centroids

31 What is the main reason to scale features (normalize/standardize) before running K-Means?

A. To prevent features with large magnitudes from dominating distance calculations
B. To convert categorical variables to numeric
C. To increase the number of clusters
D. To make the data normally distributed

32 In the context of Association Rules, the 'Consequent' is found on which side of the arrow?

A. Left (IF side)
B. Neither side
C. Right (THEN side)
D. Both sides

33 Which metric would you look at to determine if a high-confidence rule is merely a coincidence because the consequent is very popular?

A. Confidence
B. Support
C. Lift
D. Accuracy

34 If Support(A) = 0.4 and Support(A, B) = 0.2, what is the Confidence(A -> B)?

A. 0.5
B. 0.2
C. 0.8
D. 2.0

35 K-Means is sensitive to which of the following?

A. Outliers
B. All of the above
C. Scaling
D. Redundant features

36 Which clustering method generates a hierarchy of clusters?

A. Grid-Based Clustering
B. Density-Based Clustering
C. Hierarchical Clustering
D. Partitioning Clustering (K-Means)

37 In the Silhouette Analysis for K-Means, a score close to +1 indicates:

A. The point is an outlier
B. The point is assigned to the wrong cluster
C. The point is well-matched to its own cluster and far from neighboring clusters
D. The point is overlapping with other clusters

38 Which of the following scenarios is ideal for K-Means clustering?

A. Clusters are of varying densities
B. Clusters contain a lot of noise and outliers
C. Clusters are non-spherical and have irregular shapes
D. Clusters are spherical and distinct

39 In Association Rule Mining, a 'Frequent Itemset' is an itemset whose support is:

A. Less than a minimum support threshold
B. Equal to 1
C. Greater than the confidence threshold
D. Greater than or equal to a minimum support threshold

40 Which distance metric is most commonly used in K-Means?

A. Cosine Similarity
B. Hamming Distance
C. Euclidean Distance
D. Manhattan Distance

41 Which linkage method in Hierarchical Clustering aims to minimize the variance within clusters being merged (similar to K-Means)?

A. Complete Linkage
B. Single Linkage
C. Average Linkage
D. Ward's Method

42 What is a 'hard' clustering assignment?

A. The clustering is performed on hardware
B. A data point belongs to multiple clusters with varying probabilities
C. The algorithm is hard to implement
D. A data point belongs to exactly one cluster

43 In Market Basket Analysis, if {Milk, Bread} -> {Eggs} has a confidence of 0.7, it means:

A. 70% of all transactions contain Eggs
B. Eggs are bought 70% more often with Milk and Bread than expected
C. 70% of customers buy Milk and Bread
D. 70% of transactions containing Milk and Bread also contain Eggs

44 Which of the following is an application of Clustering?

A. Predicting house prices
B. Predicting credit default
C. Customer Segmentation
D. Classifying emails as spam or not spam

45 In the context of K-Means, what is a Centroid?

A. The outlier point in a cluster
B. The boundary of the cluster
C. The point closest to the origin
D. The arithmetic mean position of all the points in the cluster

46 Which of the following is true regarding the computational complexity of Hierarchical Clustering compared to K-Means for large datasets?

A. Hierarchical cannot run on large datasets
B. Hierarchical is generally slower and more memory intensive
C. They have the exact same complexity
D. Hierarchical is generally faster

47 When performing K-Means, if you initialize centroids to the same location, what happens?

A. The algorithm converges in one step
B. The algorithm fails to generate distinct clusters
C. The algorithm works perfectly
D. It automatically separates them

48 A Lift value of 0.5 suggests:

A. Negative correlation (Substitutes)
B. Independence
C. Positive correlation
D. Strong rule

49 What happens to the WCSS as the number of clusters (K) increases towards the total number of data points?

A. It decreases towards zero
B. It fluctuates randomly
C. It remains constant
D. It increases

50 Which step ensures that the K-Means algorithm converges?

A. The use of Manhattan distance
B. The fact that WCSS decreases or stays the same with every iteration
C. The randomization of initial points
D. The use of the Elbow method
Previous MCQ
Unit 3
Next MCQ
Unit 5