Bayesian Analysis of a simple button-pressing game, in which 1 button produces a reward 80% of the time, and another button produces a reward 20& of the time.
The provided script main.py
can be used to better understand Bayesian inference in a simple two-choice environment.
A user can specify any number of hypotheses about the behavior of the two choices, and run the simulation for any number of steps.
If the correct hypothesis is among the tested hypotheses, the posterior assigned to the correct hypothesis should converge to 1.0
The priors are automatically set to be equi-probable.
Developed using Python 3.7 and Ubuntu 18.04