GWKokab documentation

GWKokab offers a robust suite of probability density function (PDF) models for analyzing the distribution of parameters in compact binary coalescence (CBC) sources, such as neutron star or black hole mergers.

Researchers use these models to create synthetic populations of CBC sources, aiding in population inference studies. This helps scientists draw statistical conclusions about the astrophysical population of CBC events based on observed data. The models cover various aspects of CBC sources, including mass distribution, spin, and redshift.

Built with Distribution objects, these models leverage NumPyro’s integration with JAX for automatic differentiation and GPU acceleration, ensuring mathematical rigor and computational efficiency for large-scale simulations and analyses.

GWKokab also incorporates flowMC, a normalizing flow-enhanced sampling package for probabilistic inference, providing a powerful framework for Bayesian parameter estimation from complex, high-dimensional distributions.

The package includes a wide range of well-established gwkokab.models frequently cited in the literature. It is continuously updated with new models to keep pace with advancements in the field, and community contributions are encouraged.

GWKokab supports research in gravitational wave astronomy, offering tools for detailed parameter estimation of single events and large-scale population studies. Its flexibility and extensibility make it invaluable for researchers at all levels.

Table of Contents

Getting started

• Installation
  • Virtual Environment Setup
  • Stable Release
  • Nightly Build
• Frequently Asked Questions
  • How to write my own population model?
  • GWKokab is slow on GPU
• Citing GWKokab
• Hello GWKokab
  • Generating Mock Posterior Estimates
  • Hierarchical Bayesian Inference (Discrete Method)
  • Hierarchical Bayesian Inference (Continuous Method)
  • Expected Number of Detections and Sensitivity Estimation
  • Training a Neural Network to Estimate Sensitivity

Further resources

• Developer Documentation 💻
  • Contributing to GWKokab
  • Building from source
  • Contributor Covenant Code of Conduct