Skip to content

johertrich/Sparse_Mixture_Models

Repository files navigation

Sparse Mixture Models inspired by ANOVA Decompositions

This repository contains the numerical examples of the paper [1]. Please cite the paper, if you use this code.

It is available at
https://doi.org/10.1553/etna_vol55s142

For questions and bug reports, please contact Johannes Hertrich (j.hertrich(at)math.tu-berlin.de).

Requirements and Usage

The code is written in Python using Tensorflow 2.3.0. Usually the code is also compatible with some other versions of Tensorflow 2.x. A rough specification of each function/script is given in the corresponding header.

The script run_example.py reproduces the examples from Section 4.1 and 4.2 and the script run_grads.py reproduces the examples from Section 4.3 of [1].

Note that the code is highly experimental, far from optimized and only sparsely commented, so use it with care.

Reference

[1] J. Hertrich, F. A. Ba and G. Steidl (2022).
Sparse Mixture Models inspired by ANOVA.
Electronic Transactions on Numerical Analysis, vol. 55, pp. 142-168.

About

Numerical experiments for the paper "Sparse Mixture Models inspired by ANOVA".

Resources

License

Stars

Watchers

Forks

Languages