Hi all, I’m trying to get the posterior predictive samples which has a poisson distribution. I understand how to successfully sample from the posterior gaussian process but not the posterior y_pred as shown below. Any thoughts on how to do this the pymc3 way? with pm.Model() as model: ℓ = pm.…

Yep, that did the trick. Thanks! This is the final solution with above incorporated: test_x = np.arange(100)[:,None] with pm.Model() as model: f_pred = gp.conditional("f_pred", test_x) y_pred = pm.Poisson("y_pred", mu=tt.square(f_pred) + 1e-6, …

Poisson likelihood over a Gaussian process

Questions

junpenglao November 13, 2019, 8:15am 4

I see, try y_pred = pm.Poisson("y_pred", mu=tt.square(f_pred) + 1e-6, shape=test_x.shape)

Topic		Replies	Views
Multivariate Gaussian Process Classification (GPC) and Prediction Questions	2	915	February 23, 2020
How to calculate log posterior of a GP over a trace in pymc3 Questions	14	2286	February 2, 2018
Using a Gaussian process at intermediate level in hierarchy Questions	3	899	May 17, 2020
Shape of posterior predictive samples of GP Questions	3	555	November 10, 2019
Sample prior predictive for Gaussian process model Questions	3	868	March 8, 2020

Poisson likelihood over a Gaussian process

Related topics