distribution-optimization-py: Optimization-Based GMM Parameter Estimation

This repository contains code and scripts for Gaussian Mixture Model (GMM) parameter estimation using optimization methods.

• Generate synthetic datasets
• Benchmark optimization algorithms
• Reproduce plots and figures presented in the paper

Table of Contents

1. Installation
2. Generating Datasets
3. Benchmarking Optimization Algorithms
4. Reproducing Plots

---

Installation

1. Clone the repository:

   git clone git+https://github.com/agh-a2s/distribution-optimization.git

2. Navigate to the cloned directory:

   cd distribution-optimization

3. Install dependencies (using Poetry):

   poetry install

---

Generating Datasets

Synthetic datasets used in the paper are stored in the results directory.
To regenerate these datasets, run:

poetry run python3 -m distribution_optimization_py.experiment

This script will produce or update synthetic datasets inside the results folder.

---

Benchmarking Optimization Algorithms

To benchmark different optimization algorithms for GMM parameter estimation, run:

poetry run python3 -m distribution_optimization_py.solver

---

Reproducing Plots

All plots are generated and saved to the images directory. The scripts below reproduce the figures from the paper:

• Fig. 1:

  poetry run python3 -m distribution_optimization_py.plot_binning_scheme_difference

• Fig. 2:

  poetry run python3 -m distribution_optimization_py.experiment.plot

• Fig. 3:

  poetry run python3 -m distribution_optimization_py.solver.plot

• Fig. 4:

  poetry run python3 -m distribution_optimization_py.compare_results