Institutional Sponsor

  • Black Twitter Icon

© 2019 by SciPy.

Ralph Smith
Paul Miles

Parameter Estimation Using the Python Package pymcmcstat

Metropolis algorithms have greatly expanded our ability to estimate parameter distributions. In this talk we introduce pymcmcstat [Miles, 2018], which utilizes the Delayed Rejection Adaptive Metropolis (DRAM) algorithm [Haario et al., 2006, Haario et al., 2001] to perform Markov Chain Monte Carlo (MCMC) simulations. The user interface provides a straight forward environment for experienced and new Python users to quickly compare their models with data. Furthermore, the package provides a wide variety of diagnostic tools for visualizing uncertainty propagation. This package has been utilized in a wide array of scientific and engineering problems, including radiation source localization and constitutive model development of smart material systems.