Modeling dense relational data 
Tue Herlau, Morten Mørup, Mikkel N. Schmidt, Lars K. Hansen

Abstract  Relational modelling classically consider sparse and discrete data. Measures of influence computed pairwise between temporal sources naturally give rise to dense continuousvalued matrices, for instance pvalues from Granger causality. Due to asymmetry or lack of positive definiteness they are not naturally suited for kernel Kmeans. We propose a generative Bayesian model for dense matrices which generalize kernel Kmeans to consider offdiagonal interactions in matrices of interactions, and demonstrate its ability to detect structure on both artificial data and two real data sets. 
Type  Conference paper [With referee] 
Conference  Machine Learning for Signal Processing (MLSP), 2012 IEEE International Workshop on 
Year  2012 
BibTeX data  [bibtex] 
IMM Group(s)  Intelligent Signal Processing 