GLLiM: a flexible Matlab toolbox for Gaussian Locally Linear Mapping

The GLLiM toolbox provides a set of Matlab functions allowing to learn a relationship between two spaces. It implements the  hGLLiM algorithm described in detail in: A. Deleforge, F. Forbes, and R. Horaud. High-Dimensional Regression with Gaussian Mixtures and Partially-Latent Response Variables. Statistics and Computing. 2014. The article is available online on arXiv at

The toolbox also contains an example of application (EXAMPLE.m) using the freely available Stanford dataset for faces:

Download the toolbox

For any questions regarding the toolbox, please contact Antoine Deleforge.