pmrf.fitting

The fitting module, with built-in fitting algorithms and results.

All fitters accept the same base initialization arguments as BaseFitter. Additionally, fitters are categorized into either FrequentistFitter or BayesianFitter, and accept additional arguments accordingly.

To run the fits, several fitting routines are provided. The simplest is the fit method, which runs a single fit with the specified features and measured data. Key-word arguments are forwarded to the specified fitter’s run routine, which may provide additional run-time arguments, and also generally forwards arguments to its specific backend (for example, PolyChord or SciPy).

Fitters

`pmrf.fitting.NumPyroMCMCFitter`(model, *[, ...])	NumPyro MCMC: Markov Chain Monte Carlo (MCMC) sampling using `numpyro.infer.MCMC`.
`pmrf.fitting.NumPyroNSFitter`(model, *[, ...])	NumPyro NS: Nested sampling using `numpyro.contrib.nested_sampling.NestedSampler`.
`pmrf.fitting.SciPyMinimizeFitter`(model, *[, ...])	SciPy Minimize: Classical optimization using `scipy.optimize.minimize`.

Results

`pmrf.fitting.NumPyroResults`([measured, ...])	NumPyro: Results from `numpyro.infer`.
`pmrf.fitting.SciPyResults`([measured, ...])	SciPy: Results from `scipy.optimize`.

Bases

`pmrf.fitting.BaseFitter`(model, *[, ...])	An abstract base class that provides the foundation for all fitting algorithms in pmrf.
`pmrf.fitting.FrequentistFitter`(model, *[, ...])	A base class for frequentist (classical) optimization methods.
`pmrf.fitting.BayesianFitter`(model, *[, ...])	A base class for Bayesian fitting methods.
`pmrf.fitting.FitResults`([measured, ...])	Container for the results of a model fitting process.
`pmrf.fitting.FrequentistResults`([measured, ...])	Results obtained from a Frequentist (classical) fitting process.
`pmrf.fitting.BayesianResults`([measured, ...])	Abstract base class for results obtained from a Bayesian fitting process.
`pmrf.fitting.FitContext`(model, measured, ...)	Context object holding the state and data required for a fit execution.
`pmrf.fitting.FrequentistContext`(model, ...)	Context object for Frequentist fitting, containing the cost function.
`pmrf.fitting.BayesianContext`(model, ...[, ...])	Context object for Bayesian fitting, containing likelihood configurations.