Details about ROXY
| Category | Information |
|---|---|
| Purpose | ROXY (Regression and Optimisation with X and Y errors) is a python package for performing MCMC where the data have both x and y errors. The common approach for this problem is to use a Gaussian likelihood with a mean given by f(x_{obs}, \theta) and a variance \sigma_y^2 + f'(x_{obs}, \theta)^2 \sigma_x^2, but this ignores the underlying distribution of true x values and thus gives biased results. Instead, this package allows one to use the MNR (Marginalised Normal Regression) method which does not exhibit such biases. |
| Related community |
Action Dark Energy
|
| Tool type | Software |
| URL | https://github.com/DeaglanBartlett/roxy |
| Documentation | https://roxy.readthedocs.io/en/latest/?badge=latest |
| Author | Deaglan Bartlett |
| Publication: | |
| Language | Python 3.x |
| Keywords |
|
Comments: