Machine Learning Clustering 2 — Questions and Answers
Question 1: Which linkage criterion in hierarchical clustering merges clusters based on the maximum distance between any two points in different clusters?
- Single linkage
- Complete linkage (Correct answer)
- Average linkage
- Ward linkage
Correct answer: Complete linkage
Complete linkage uses the maximum pairwise distance between clusters, producing compact, spherical clusters.
Question 2: What is the primary disadvantage of the DBSCAN algorithm?
- It cannot handle noise
- It requires specifying K in advance
- It struggles with clusters of varying density (Correct answer)
- It only works on linear data
Correct answer: It struggles with clusters of varying density
DBSCAN uses a fixed epsilon and minPts, making it sensitive to clusters with significantly different densities.
Question 3: In the Expectation-Maximization (EM) algorithm for Gaussian Mixture Models, what does the E-step compute?
- New cluster centroids
- The covariance matrices
- The posterior probability of each point belonging to each component (Correct answer)
- The number of clusters
Correct answer: The posterior probability of each point belonging to each component
The E-step computes the responsibility (posterior probability) of each Gaussian component for every data point.
Question 4: Which internal clustering evaluation metric measures how similar a point is to its own cluster compared to other clusters?
- Davies-Bouldin Index
- Silhouette Score (Correct answer)
- Calinski-Harabasz Index
- Inertia
Correct answer: Silhouette Score
The Silhouette Score ranges from -1 to 1 and compares intra-cluster cohesion to inter-cluster separation for each point.
Question 5: What does 'inertia' measure in the context of K-Means clustering?
- The number of iterations until convergence
- The sum of squared distances from each point to its assigned centroid (Correct answer)
- The ratio of between-cluster to within-cluster variance
- The average silhouette score across all points
Correct answer: The sum of squared distances from each point to its assigned centroid
Inertia (within-cluster sum of squares) measures the compactness of clusters; lower inertia indicates tighter clusters.
Question 6: Which of the following clustering algorithms is specifically designed to handle arbitrarily shaped clusters?
- K-Means
- Gaussian Mixture Models
- DBSCAN (Correct answer)
- Hierarchical Agglomerative Clustering with Ward linkage
Correct answer: DBSCAN
DBSCAN groups points based on density connectivity, allowing it to find clusters of any shape unlike centroid-based methods.
Question 7: What is a dendrogram used for in hierarchical clustering?
- To plot the decision boundary between clusters
- To visualize the merging sequence and distances between clusters (Correct answer)
- To show the silhouette score for each cluster
- To select the optimal epsilon for DBSCAN
Correct answer: To visualize the merging sequence and distances between clusters
A dendrogram is a tree diagram that records the sequence and distance at which clusters are merged, helping select the number of clusters.
Machine Learning Clustering 2
Which linkage criterion in hierarchical clustering merges clusters based on the maximum distance between any two points in different clusters?