Prior¶
The pybalonor package supports a prior that is uniform in the mode of logs \(l_0\) and the standard deviation of logs \(l_1\),
\[\phi(l_0, l_1 \,|\, l_0^\text{min}, l_0^\text{max}, l_1^\text{min},
l_1^\text{max}) = \frac{1}{(l_0^\text{max} - l_0^\text{min})
(l_1^\text{max} - l_1^\text{min})}\,,\]
where the rectangular support of \(l_0\) and \(l_1\) is defined by
\[l_0^\text{min}>0\,, \quad l_0^\text{max}>l_0^\text{min},\quad
l_1^\text{min}>0\,,\quad \text{and} \quad l_1^\text{max}>l_1^\text{min}\,.\]
The uniform prior weights all parameter combinations \((l_0,l_1)\) within the bounds
\[ \begin{align}\begin{aligned}l_0^\text{min} \leq l_0 \leq l_0^\text{max}\\l_1^\text{min} \leq l_1 \leq l_1^\text{max}\end{aligned}\end{align} \]
equally.
FlatPrior
¶
- class FlatPrior(l0_min, l0_max, l1_min, l1_max)¶
Flat prior of the log-normal distribution.
- Parameters
l0_min (float) – Minimum \(l_0\) within the prior support.
l0_max (float) – Maximum \(l_0\) within the prior support. Has to fulfill \(\,l_0^\text{max} > l_0^\text{min}\).
l1_min (float) – Minimum \(l_1\) within the prior support. Has to fulfill \(\,l_1^\text{min} \geq 0\).
l1_max (float) – Maximum \(l_1\) within the prior support. Has to fulfill \(\,l_1^\text{max} > l_1^\text{min}\).