Clustering Partially Observed Graphs via Convex Optimization

Abstract

This paper considers the problem of clustering a partially observed unweighted graph---i.e., one where for some node pairs we know there is an edge between them, for some others we know there is no edge, and for the remaining we do not know whether or not there is an edge. We want to organize the nodes into disjoint clusters so that there is relatively dense (observed) connectivity within clusters, and sparse across clusters. We take a novel yet natural approach to this problem, by focusing on finding the clustering that minimizes the number of "disagreements"---i.e., the sum of the number of (observed) missing edges within clusters, and (observed) present edges across clusters. Our algorithm uses convex optimization; its basis is a reduction of disagreement minimization to the problem of recovering an (unknown) low-rank matrix and an (unknown) sparse matrix from their partially observed sum. We evaluate the performance of our algorithm on the classical Planted Partition/Stochastic Block Model. Our main theorem provides sufficient conditions for the success of our algorithm as a function of the minimum cluster size, edge density and observation probability; in particular, the results characterize the tradeoff between the observation probability and the edge density gap. When there are a constant number of clusters of equal size, our results are optimal up to logarithmic factors.

This is the final version published in Journal of Machine Learning Research (JMLR). Partial results appeared in International Conference on Machine Learning (ICML) 2011

References in corpus (5)

• The Augmented Lagrange Multiplier Method for Exact Recovery of Corrupted Low-Rank Matrices
• Asymptotic analysis of the stochastic block model for modular networks and its algorithmic applications
• Finding Dense Clusters via "Low Rank + Sparse" Decomposition
• Performance Analysis of Spectral Clustering on Compressed, Incomplete and Inaccurate Measurements
• Efficient Clustering with Limited Distance Information

Cited by in corpus (26)

• Statistical-Computational Tradeoffs in Planted Problems and Submatrix Localization with a Growing Number of Clusters and Submatrices
• Near-Optimal Joint Object Matching via Convex Relaxation
• Semi-Blind Inference of Topologies and Dynamical Processes over Graphs
• Guaranteed clustering and biclustering via semidefinite programming
• Convexity in source separation: Models, geometry, and algorithms
• Noisy Sparse Subspace Clustering
• Achieving Exact Cluster Recovery Threshold via Semidefinite Programming: Extensions
• The achievable performance of convex demixing
• Clustering with Noisy Queries
• Breaking the Small Cluster Barrier of Graph Clustering
• Scalable and Robust Community Detection with Randomized Sketching
• Classification from Pairwise Similarity and Unlabeled Data
• Mixed Membership Graph Clustering via Systematic Edge Query
• Subspace Estimation from Unbalanced and Incomplete Data Matrices: ${\ell }_{2,\infty }$ Statistical Guarantees
• Learning from Untrusted Data
• Correlation Clustering with Noisy Partial Information
• Learning Social Circles in Ego Networks based on Multi-View Social Graphs
• Clustering with Fast, Automated and Reproducible assessment applied to longitudinal neural tracking
• Efficient Online Minimization for Low-Rank Subspace Clustering
• Matrix Completion with Prior Subspace Information via Maximizing Correlation
• Co-factor analysis of citation networks
• Robust Correlation Clustering with Asymmetric Noise
• Robust Clustering Oracle and Local Reconstructor of Cluster Structure of Graphs
• Clustering Degree-Corrected Stochastic Block Model with Outliers
• Low Rank Regularization: A Review
• K-median: exact recovery in the extended stochastic ball model