I have been struggling with interpreting the Switch Operator in PYMC.
I have a given time-series in which I presume to have 2 Switch points.
By taking the Radon example provided in PYMC docs., I have implemented a similar model in order to evaluate my posterior distribution in respect to Y.
Also, in order to account the switch Operator to be visible by the PYMC model, I had to convert those switches to be deterministic variables by means of the pm.Deterministc method.
After this conversion, I can visualize the all prior distributions and the switch points. Nevertheless, the interpretation of the switch points become clumsy due to the multiple histogram plots.
Therefore, can anyone provide some example and discussion of how one can interpret a switch Operator distribution?