Hi, I am using the marginal likelihood implementation of PyMC3 GP. How can I calculate the marginal likelihood for model comparison? My understanding is that model.logp(y) would give me the log posterior. A bit unrelated, I tried calculating MLE for the GP model using flat priors and pm.find_MAP() …

The length scale and noise only takes positive value - you should use HalfFlat instead. noise_obs = pm.HalfFlat('noise_obs') lengthscale = pm.HalfFlat('lengthscale')

Calculating marginal likelihood of GP Model

Questions

junpenglao December 31, 2018, 12:02am 7

You can do model.logp(mp) to get the model logp (up to a constant) at the point of MLE.

Topic		Replies	Views
Maximum likelihood estimation of a Bayesian model Questions	11	5575	April 12, 2018
Kernel hyper-parameters priors and their use in MAP estimates of Gaussian Processes Questions	6	1290	May 18, 2022
Unconsistent logp, Evaluating the results of find_MAP on Custom Likelihood Questions	3	610	August 28, 2018
What makes difference find_MAP() and pm.sample() Questions	2	2422	October 17, 2021
GP and MC integral approximation - which logp to use? Questions	2	576	February 12, 2019

Calculating marginal likelihood of GP Model

Related topics