Which density-based method defines similarity by overlap in nearest neighbors rather than raw distance?

Which density-based method defines similarity by overlap in nearest neighbors rather than raw distance?

• A. Density-based Clustering
• B. SNN (Shared Nearest Neighbors)
• C. DBSCAN (Density-Based Spatial Clustering of Applications with Noise)
• D. Agglomerative Clustering

Answer: (B)

This item tests defining similarity by how much two points share in their closest neighbors rather than by their direct distance. In the Shared Nearest Neighbors approach, you first find the k nearest neighbors for each point. Then the similarity between two points is based on the overlap of their neighbor sets, such as the size of the intersection |N_k(i) ∩ N_k(j)| (often with normalization). A larger overlap signals that the two points inhabit similar local contexts, even if their direct distance isn’t small. Clustering proceeds by linking points with high shared-neighbor similarity, forming groups that reflect local structure and are more robust to varying densities than methods relying solely on raw distance. Other density- or distance-based methods rely on distance thresholds or density reachability using raw distances, which can struggle when densities vary or when absolute distances don’t capture local relationships.

This item tests defining similarity by how much two points share in their closest neighbors rather than by their direct distance. In the Shared Nearest Neighbors approach, you first find the k nearest neighbors for each point. Then the similarity between two points is based on the overlap of their neighbor sets, such as the size of the intersection |N_k(i) ∩ N_k(j)| (often with normalization). A larger overlap signals that the two points inhabit similar local contexts, even if their direct distance isn’t small. Clustering proceeds by linking points with high shared-neighbor similarity, forming groups that reflect local structure and are more robust to varying densities than methods relying solely on raw distance.

Other density- or distance-based methods rely on distance thresholds or density reachability using raw distances, which can struggle when densities vary or when absolute distances don’t capture local relationships.