Regression with variable known uncertainty with GPFlow

Question

I am trying to model some 1D data where each data point has a different error bar.

In the GPFlow documentation I find an example of how to do this here (known noise variances demo)

However, the demo doesn't seem to produce a GP model that models the data well and I was wondering if anyone had any suggests to improve the GP fit in the demo?

Answer 1

This was fixed in this PR. It was a typo in the documentation.