Online Program Home
My Program

Abstract Details

Activity Number: 81
Type: Contributed
Date/Time: Sunday, July 31, 2016 : 4:00 PM to 5:50 PM
Sponsor: ENAR
Abstract #320997
Title: Overlapping Community Detection in Networks via Sparse Principal Component Analysis
Author(s): Jesus Daniel Arroyo Relion* and Elizaveta Levina
Companies: University of Michigan and University of Michigan
Keywords: community detection ; sparse PCA ; networks ; overlapping communities

Community detection in networks, the problem of finding groups of nodes that have more connections to each other than to the rest of the network, has received a lot of attention in the literature, but many methods only allow for a node to belong to exactly one community. In practice, nodes in a network may belong to multiple communities. Here we propose a new efficient algorithm for overlapping community detection based on sparse principal component analysis. The algorithm has a computational cost similar to that of estimating the largest eigenvectors of the adjacency matrix, and does not require an additional clustering step like spectral clustering techniques. We show that our method is consistent in selecting the community memberships under an overlapping version of the stochastic blockmodel and evaluate the method empirically on simulated and real-world networks, showing good statistical performance and computational efficiency.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association