MS-DS Master of Data science FREE MS-DS Master of Data science Unsupervised Machine Learning Models Questions and Answers 2 — Questions and Answers
Question 1: In DBSCAN, what is the role of the epsilon (eps) parameter?
- It defines the maximum distance between two points to be considered neighbors (Correct answer)
- It sets the minimum number of clusters
- It determines the learning rate of the algorithm
- It specifies the dimensionality of the feature space
Correct answer: It defines the maximum distance between two points to be considered neighbors
The epsilon parameter in DBSCAN defines the radius of the neighborhood around each point used to identify density-connected regions.
Question 2: Which technique is most appropriate for reducing high-dimensional data while preserving local neighborhood structure?
- Principal Component Analysis
- t-Distributed Stochastic Neighbor Embedding (t-SNE) (Correct answer)
- Linear Discriminant Analysis
- Variance Inflation Factor analysis
Correct answer: t-Distributed Stochastic Neighbor Embedding (t-SNE)
t-SNE is specifically designed to preserve local neighborhood relationships when projecting high-dimensional data into lower dimensions for visualization.
Question 3: What does the silhouette coefficient measure in cluster analysis?
- How similar an object is to its own cluster compared to other clusters (Correct answer)
- The total variance explained by the clustering
- The number of optimal clusters
- The distance between cluster centroids only
Correct answer: How similar an object is to its own cluster compared to other clusters
The silhouette coefficient compares intra-cluster cohesion with inter-cluster separation, ranging from -1 to 1 where higher values indicate better-defined clusters.
Question 4: In Gaussian Mixture Models (GMM), what algorithm is typically used to estimate the model parameters?
- Gradient Descent
- Expectation-Maximization (EM) (Correct answer)
- Stochastic Gradient Descent
- Newton-Raphson method
Correct answer: Expectation-Maximization (EM)
The Expectation-Maximization algorithm iteratively estimates the posterior probabilities of cluster memberships (E-step) and updates the Gaussian parameters (M-step) until convergence.
Question 5: What is a key advantage of hierarchical agglomerative clustering over K-Means?
- It always runs faster on large datasets
- It does not require specifying the number of clusters in advance (Correct answer)
- It guarantees globally optimal clusters
- It only works with numerical features
Correct answer: It does not require specifying the number of clusters in advance
Hierarchical agglomerative clustering builds a dendrogram that can be cut at any level, allowing the analyst to choose the number of clusters after the algorithm has run.
Question 6: Which statement best describes the concept of reconstruction error in autoencoders used for unsupervised learning?
- The difference between the input data and the output produced by the decoder (Correct answer)
- The number of neurons lost during dropout
- The gradient magnitude during backpropagation
- The ratio of training samples to test samples
Correct answer: The difference between the input data and the output produced by the decoder
Reconstruction error quantifies how well the autoencoder can reproduce its input after compressing it through the bottleneck layer, serving as the primary loss function.
MS-DS Master of Data science FREE MS-DS Master of Data science Unsupervised Machine Learning Models Questions and Answers 2
In DBSCAN, what is the role of the epsilon (eps) parameter?