I am trying to model a process, which looks a bit like a function, with heavy-tailed noise, and some values are also multiplied by a random value between 0 and 1. Consider the following example data. rng = np.random.default_rng(4) x = np.linspace(-3, 3, 50) y = 5 + np.sin(x) noise = 0.2 * rng.stand…

The approach with clip is a neat idea, but unfortunately this is makes the posterior as seen by the sampler quite tricky. If the eff parameter is larger than 1 the derivative will be zero, and the sampler can’t know that making it smaller will eventually change the logp. In general it is good to avo…

Gaussian Processes and Unusual Noise

Questions v5

cluhmann April 28, 2023, 3:23am 4

There are definitely examples that people have posted here (e.g., this old example), but I’m not sure there is a clean, simple example in the notebook gallery/documentation.

1 Like

Topic		Replies	Views
Latent gaussian process prediction problem Questions gaussian_process	3	1037	January 25, 2022
Gaussian process regression: samples from mean model Questions	1	376	August 25, 2020
Gaussian process using noisy data training points Questions gaussian_process	11	1138	February 13, 2022
Latent Gaussian Processes with input uncertainties Questions	2	507	May 22, 2018
Gaussian Process -Statistical rethinking v5 gaussian_process	2	810	January 28, 2023

Gaussian Processes and Unusual Noise

Related topics