This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: constant values in 2 nd eigenvector 4/6/10 COT 6936 4 Graph Bisection (Continued) Eigenvalues indicate strength of bisection How to get bisections with n/2 vertices? Use median value in second eigenvector How to get k partitions? Perform bisections recursively Use more eigenvectors 4/6/10 COT 6936 5 Spectral Clustering: Strategy Given data points and a distance function, construct a weighted graph Let A be its adjacency matrix; let D be diagonal matrix with degrees along diagonal Construct Laplacian L ( PSD , nonneg eigenv .) Unnormalized: L = D A Normalized symmetric: L = D1/2 LD 1/2 Random Walk: L = D1 L Matrix L k has cols = first k eigenvectors of L Cluster rows of L k . 4/6/10 COT 6936 6 3 Spectral Clustering Need distance measure (need not be a metric), i.e., triangle inequality not needed Not Modelbased Global method Turns discrete problem into continuous 4/6/10 COT 6936 7...
View Full
Document
 Spring '12
 Giri
 Algorithms

Click to edit the document details