The repository contains MATLAB code for the VB-WMM model. The model aims to cluster a set of Gram matrices originating from zero-mean Gaussian data. Coordinate ascent variational inference is used, and due to the chosen priors and Q-distribution all updates are closed form achieved by moment matching.
Please see the following publication for further details on the model and its application to dynamic functional brain connecitivity,
Nielsen, S. F. V., Madsen, K. H., Schmidt, M. N., & Mørup, M. (2017) Modeling Dynamic Functional Connectivity using a Wishart Mixture Model. In 2017 International Workshop on Pattern Recognition in NeuroImaging (PRNI), IEEE
MIT License (see LICENSE file for further information)
Code was developed as part of the "Brain Connectivity" project carried out at Technical University of Denmark https://brainconnectivity.compute.dtu.dk/