The Canonical Polyadic (CP) decomposition with non-negativity constraints was used on gene expression data from human brains (cf. Allen Brain Atlas) in the publication Non-negative Tensor Factorization with Missing Data for the Modeling of Gene Expressions in the Human Brain. The implementation in MATLAB was optimized for missing data entries and can be seen and downloaded from the link below.
In a project carried out together with Jesper L. Hinrich and Julian K. Larsen, supervised by Mikkel N. Schmidt, a hierarchical non-parametric Bayesian model (cf. Schmidt et. al. (2014) ) for undirected network data was implemented in C++.
Non-parametric Dynamic Functional Connectivity Modeling (NDFC)
The NDFC software was used to model dynamic functional connectivity in fMRI and EEG data using a non-parametric Hidden Markov Model (Beal 2002). The code is based on Juergen Van Gaels IHMM-Toolbox .