From 722a3dfbf82f7d8fb05b6d84c3a72fecaec5e391 Mon Sep 17 00:00:00 2001 From: Daniel Weindl Date: Thu, 11 Dec 2025 08:58:10 +0100 Subject: [PATCH 1/2] Clarify prior & bounds Make the parameter prior description more consistent and explicit. Previously, it said that parameter bounds truncate the prior distributions, but the provided PDFs suggested otherwise. --- doc/v2/documentation_data_format.rst | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/doc/v2/documentation_data_format.rst b/doc/v2/documentation_data_format.rst index 0365154d..e43aa040 100644 --- a/doc/v2/documentation_data_format.rst +++ b/doc/v2/documentation_data_format.rst @@ -916,6 +916,20 @@ Detailed field description Prior distributions ~~~~~~~~~~~~~~~~~~~ +The prior distributions supported for the +:ref:`MAP objective function and for Bayesian inference ` +are listed below. PEtab only supports univariate prior distributions. +The probability density functions (PDFs) below assume that the parameter bounds +are wide enough to not truncate the distributions. If the parameter bounds +are narrower than the distribution's support, the distributions are truncated, +resulting in the following effective prior distribution: + +.. math:: + \pi_{\text{eff}}(x) = \frac{\pi(x)}{\text{CDF}(\text{upperBound}) - \text{CDF}(\text{lowerBound})} + +where :math:`\pi(x)` is the PDF of the non-truncated distribution +and :math:`\text{CDF}(\cdot)` its cumulative distribution function. + Let :math:`x` denote the parameter value and :math:`\Gamma` the `Gamma function `__., then the following prior distributions are supported: From 63bc8ba3494439f9791515c340dc5281e3f9a20e Mon Sep 17 00:00:00 2001 From: Daniel Weindl Date: Fri, 12 Dec 2025 08:12:36 +0100 Subject: [PATCH 2/2] Update doc/v2/documentation_data_format.rst Co-authored-by: Dilan Pathirana <59329744+dilpath@users.noreply.github.com> --- doc/v2/documentation_data_format.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/doc/v2/documentation_data_format.rst b/doc/v2/documentation_data_format.rst index e43aa040..064b4f02 100644 --- a/doc/v2/documentation_data_format.rst +++ b/doc/v2/documentation_data_format.rst @@ -922,10 +922,10 @@ are listed below. PEtab only supports univariate prior distributions. The probability density functions (PDFs) below assume that the parameter bounds are wide enough to not truncate the distributions. If the parameter bounds are narrower than the distribution's support, the distributions are truncated, -resulting in the following effective prior distribution: +resulting in the following truncated prior distribution: .. math:: - \pi_{\text{eff}}(x) = \frac{\pi(x)}{\text{CDF}(\text{upperBound}) - \text{CDF}(\text{lowerBound})} + \pi_{\text{trunc}}(x) = \frac{\pi(x)}{\text{CDF}(\text{upperBound}) - \text{CDF}(\text{lowerBound})} where :math:`\pi(x)` is the PDF of the non-truncated distribution and :math:`\text{CDF}(\cdot)` its cumulative distribution function.