GATE DA Machine Learning – Full Mock Test-1 ⋆ CSE Study247

Question 56

MCQ · Single Correct 1 mark

In Hierarchical Agglomerative Clustering with N samples, how many total merge steps are performed to build a complete dendrogram ending with 1 cluster?

A log2(N) B N C $N^2$ D N - 1

Question 57

MCQ · Single Correct 2 mark

What is the Dunn Index for evaluating cluster quality?

A Ratio of average variance to total variance B Log likelihood of cluster assignment C Ratio of minimum inter-cluster distance to maximum intra-cluster diameter (higher is better) D Difference between between-cluster and within-cluster sum of squares

Question 58

MSQ · Multiple Correct 2 mark

Which of the following statements regarding PRINCIPAL COMPONENT ANALYSIS (PCA) are CORRECT?

A Principal components are orthogonal to each other. B PCA minimizes the sum of squared reconstruction errors. C Total variance of the data is equal to the sum of eigenvalues of the covariance matrix. D PCA requires target class labels y to compute component axes.

Question 59

MSQ · Multiple Correct 2 mark

Which of the following are INTRINSIC (unsupervised) cluster validity indices?

A Silhouette Coefficient B Davies-Bouldin Index C Dunn Index D Adjusted Rand Index (ARI)

Question 60

MSQ · Multiple Correct 2 mark

Which of the following statements comparing k-Means and DBSCAN are CORRECT?

A k-Means requires specifying the number of clusters k in advance, whereas DBSCAN does not. B DBSCAN can find arbitrarily shaped clusters, whereas k-Means assumes convex/spherical clusters. C DBSCAN explicitly identifies noise points, whereas k-Means assigns every point to a cluster. D k-Means is sensitive to initial centroid selection, whereas DBSCAN results are deterministic for core points.