Fast link prediction for large networks using spectral embedding. (26th July 2017)

• Record Type: Journal Article
• Authors: Pachev, Benjamin; Webb, Benjamin
• Is Part Of: Journal of complex networks. Volume 6:Number 1(2018)
• Page Start: 79
• Page End: 94
• Publication Date: 2017-07-26
• Subjects: link prediction -- graph embedding -- commute time -- resistance distance -- closest pairs; Numerical analysis -- Periodicals; Computer networks -- Periodicals; Social networks -- Periodicals; 518.05
• DOI: 10.1093/comnet/cnx021
• Languages: English
• ISSNs: 2051-1310

Abstract

Abstract: Many link prediction algorithms require the computation of a similarity metric on each vertex pair, which is quadratic in the number of vertices and infeasible for large networks. We develop a class of link prediction algorithms based on a spectral embedding and the $k$ closest pairs algorithm that are scalable to very large networks. We compare the prediction accuracy and runtime of these methods to existing algorithms on several large link prediction tasks. Our methods achieve comparable accuracy to standard algorithms but are significantly faster.