Conference Proceedings

Semi-supervised Blockmodelling with Pairwise Guidance

Mohadeseh Ganji, Jeffrey Chan, Peter J Stuckey, James Bailey, Christopher Leckie, Kotagiri Ramamohanarao, Laurence Park, M Berlingerio (ed.), F Bonchi (ed.), T Gartner (ed.), N Hurley (ed.), G Ifrim (ed.)

MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, ECML PKDD 2018, PT II | SPRINGER INTERNATIONAL PUBLISHING AG | Published : 2019

Abstract

Blockmodelling is an important technique for detecting underlying patterns in graphs. Existing blockmodelling algorithms are unsupervised and cannot take advantage of the existing information that might be available about objects that are known to be similar. This background information can help finding complex patterns, such as hierarchical or ring blockmodel structures, which are difficult for traditional blockmodelling algorithms to detect. In this paper, we propose a new semi-supervised framework for blockmodelling, which allows background information to be incorporated in the form of pairwise membership information. Our proposed framework is based on the use of Lagrange multipliers and ..

View full abstract