Statistics#

Set of elementary examples showing how the pyoed.stats subpackage can be employed.

Distributions #

Driver script to test Sampling various algorithms with simple PDFs.

Functions

`main`([run_mcmc, run_hmc, run_rejection])	Main driver.
`banana_potential_energy_value`(state[, a, b, rho])	Potential energy of the posterir.
`banana_potential_energy_gradient`(state[, a, ...])	Gradient of the Potential energy of the posterir.
`banana_log_density`(state[, a, b, rho])	The logarithm of the banana distribution PDF
`banana_log_density_gradient`(state[, a, b, rho])	The gradient of the logarithm of the banana distribution PDF
`mcmc_sample_banana_distribution`([...])
`rejection_sample_banana_distribution`([...])

get_base_output_dir()[source]#: Base output dir that enables modifying base output dir by changing pyoed settings class This is very helpful to globally steer output e.g., when running unittests.

banana_potential_energy_value(state, a=2.15, b=0.75, rho=0.9)[source]#: Potential energy of the posterir. This is dependent on the target state, not the momentum. It is the negative the posterior-log, and MUST be implemented for each distribution

banana_potential_energy_gradient(state, a=2.15, b=0.75, rho=0.9)[source]#: Gradient of the Potential energy of the posterir.

banana_log_density(state, a=2.15, b=0.75, rho=0.9)[source]#: The logarithm of the banana distribution PDF

banana_log_density_gradient(state, a=2.15, b=0.75, rho=0.9)[source]#: The gradient of the logarithm of the banana distribution PDF

mcmc_sample_banana_distribution(sample_size=1000, method='hmc')[source]#

rejection_sample_banana_distribution(sample_size=1000, proposal_mean=[0, 3], proposal_variance=2)[source]#

main(run_mcmc=True, run_hmc=True, run_rejection=True)[source]#

Main driver. Run all the sampler examples implemented in this module based on the passed flags

Parameters:

run_mcmc (bool) – if True, call/run the function mcmc_sample_banana_distribution
run_hmc (bool) – if True, call/run the function hmc_sample_banana_distribution
run_rejection (bool) – if True, call/run the function rejection_sample_banana_distribution