Performance analysis of spectral community detection in realistic graph models

Abstract : This article proposes a spectral analysis of dense random graphs generated by (a modified version of) the degree-corrected stochastic block model, for a setting where the inter block probabilities differ by O(n-) with n the number of nodes. We study a normalized version of the graph modularity matrix which is shown to be asymptotically well approximated by an analytically tractable (spiked) random matrix. The analysis of the latter allows for the precise evaluation of (i) the transition phase where clustering becomes asymptotically feasible and (ii) the alignment between the dominant eigenvectors and the block-wise canonical basis, thus enabling the estimation of misclassification rates (prior to post-processing) in simple scenarios.
Type de document :
Communication dans un congrès
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Mar 2016, Shanghai, China. 2016, 〈10.1109/icassp.2016.7472538〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01633452
Contributeur : Romain Couillet <>
Soumis le : lundi 13 novembre 2017 - 09:10:14
Dernière modification le : lundi 17 décembre 2018 - 14:28:06

Lien texte intégral

Identifiants

Citation

Hafiz Tiomoko Ali, Romain Couillet. Performance analysis of spectral community detection in realistic graph models. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Mar 2016, Shanghai, China. 2016, 〈10.1109/icassp.2016.7472538〉. 〈hal-01633452〉

Partager

Métriques

Consultations de la notice

99