How to obtain the posterior probability of 2D distribution?

Questions version agnostic
4

Some questions:

  • What does rffw.rffw do?

  • I don’t see any uncertainty in your forward model. Are rf_matrix and amp deterministic (conditional on parameters), and you are just thinking about measurement error?

  • What is the story with your custom likelihood function? It looks like two independent gaussians with \sigma=0.01 and \sigma=0.1 (and no normalizing constant). You could just have pm.Normal(mu=waveform_forward_data, sigma=0.01, observed=obs_waveform_data) and pm.Normal(mu=amp_forward_data, sigma=0.1, observed=obs_amp_data_np ).