Jump to content

User:Hosein.hashemi

From Wikipedia, the free encyclopedia


ProFit is a new publicly available code for Bayesian two-dimensional photometric galaxy profile modelling and decomposition. ProFit uses a fast C++ image generation library (libprofit) and a flexible interface to a large number of likelihood samplers. libprofit offers fast and accurate 2D integration for a useful number of profiles, including Sersic, Core-Sersic, broken-exponential, Ferrer, Moffat, empirical King, point-source and sky, with a simple mechanism for adding new profiles.

Initially ProFit comes in three varieties: a fully featured R package (ProFit), a basic PYTHON wrapper (pyprofit) and a command line terminal interface (profit-cli). However, the R package is the most advanced in terms of features and the most up to dated version.

The core libprofit library is designed to be fast and accurate at integrating and convolving target model images, where the likelihood function to compute, and optimization engine to use, are largely choices for the user via higher level interfaces. The R package version of ProFit contains a large number of examples using simple down-hill gradient schemes, more complex genetic algorithms and more computationally expensive MCMC techniques.

The core advances that ProFit offers over currently available software are that it:

• is fully open-source with multiple active developers,

• offers a standalone library (libprofit) for accurate and fast pixel integrations when generating a model,

• can be extended with new profiles in a simple well- documented manner,

• allows for simple or complex priors on parameters (an important aspect of Bayesian analysis),

• offers a range of likelihood calculations,

• is untied to any specific optimizers but has easy access to downhill minimization, genetic algorithm, and MCMC routines,

• can fit parameters in log or linear space,

• allows for simple or complex additional constraints be- tween parameters,

• offers brute-force and FFT PSF convolution options, with automatic benchmarking to select the fastest strategy.


Galaxy model Image with Boxiness
A two-components Sersic with a Ferrer bar modelled by ProFit.
Modelled Galaxy2
The approximate projected one-dimensional profile for G266033. The grey shaded region shows the 1σ error region for the data pixels.
The triangle plot of the stationary MCMC chains for the model fit of G266033 .

Libprofit model Image without convolution by the PSF

Modelled Galaxy

Modelled Galaxy by ProFit



References

[edit]

Robotham, A. S. G., Taranu, D. S., Tobar, R., Moffett, A., & Driver, S. P. (2016). ProFit: Bayesian profile fitting of galaxy images. Monthly Notices of the Royal Astronomical Society, 466(2), 1513-1541. (http://adsabs.harvard.edu/abs/2017MNRAS.466.1513R)

[edit]



Category:Astronomy Category:Physics Category:Galaxies Category:Sersic