cabouman/gmcluster

EM Clustering Algorithm

GMCluster

GMCluster: An Unsupervised Algorithm for Modeling Gaussian Mixtures

This is an EM-based clustering package for python that is based on the following C package: >https://engineering.purdue.edu/~bouman/software/cluster/

The documentation for this package is available here: >https://gmcluster.readthedocs.io/

Installing

1. Clone or download the repository and get inside:

git clone https://github.com/cabouman/gmcluster.git
cd gmcluster

2. Install the conda environment and package

   1. Option 1: Clean install from dev_scripts

      *****You can skip all other steps if you do a clean install.****

      To do a clean install, use the command:

      cd dev_scripts
      source clean_install_all.sh
      cd ..
      

   2. Option 2: Manual install

      1. Create conda environment:

         Create a new conda environment named gmcluster using the following commands:

         conda create --name gmcluster python=3.9
         conda activate gmcluster
         pip install -r requirements.txt
         

         Anytime you want to use this package, this gmcluster environment should be activated with the following:

         conda activate gmcluster
         

      2. Install gmcluster package:

         Use the following command to install the package.

         pip install .
         

         To allow editing of the package source while using the package, use

         pip install -e .
         

      3. Build the documentation:

         Use the following command to build the documentation.

         cd docs
         pip install -r requirements.txt
         make clean html
         cd ..
         

Running Demo(s)

You can validate the installation by running demo scripts.

cd demo
python demo_1.py