This is an interface to InfoNest that simplifies the process of calculating H(theta | data).
(c) 2018 Brendon J. Brewer. LICENCE: MIT.
First, clone the repository recursively:
git clone --recursive https://github.com/eggplantbren/PosteriorEntropy
Then compile the C++.
make
Run the executable:
./main
View the results (can be done while main
is still running):
python postprocess.py
The last line of output shows the measured (differential) conditional entropy H(theta | data), i.e., the expected entropy of the posterior. The example model is specified in include/Demo.h. The parameter of interest in the demo is the log width of the transit, and the prior entropy is H(theta) = 1.419 nats.