Gaussian Process extension: multiple hyperparameters #2549
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…hyperparams and columns are different features (as opposed to a list corresponding to only one [multidimensional] hyperparm). fmin_cobyla is the only optimizer that handles a matrix theta (if Welch, forces fmin_cobyla). Verbose setting outputs optimal theta.
…in checking theta for builtin correlation functions.
| y2_pred, MSE2 = gp.predict(X2, eval_MSE=True) | ||
|
|
||
| assert_true(np.allclose(y_pred, y) and np.allclose(MSE, 0.) | ||
| and np.allclose(MSE2, 0., atol=10)) |
There was a problem hiding this comment.
It's better to use assert_array_almost_equal instead of assert_true(np.allclose()) to get meaningful failure messages.
|
Maybe one of the example should be updated or extended to demonstrate this new capability too (in a similar way as done in the test). |
|
Closing. We now have a new implementation of gaussian processes (see #4270). Thanks for you contribution. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hello.
I made some simple changes to the gaussian process module. Previously, the hyperparameters theta (parameters to optimize from the kernel function) were limited to one, though that one hyperparam could be multifeatured. This is no good if you want to combine kernels (like in Rasmussen & Williams' book) because each part of the new kernel will have its own hyperparams, so I modified the code to allow this.
Theta can now be a matrix, where columns are features / dimensions (just as before), but rows now correspond to different hyperparameters.
I also added a test for it and added a check in the built-in correlation functions to make sure they aren't taking in bad parameters. All the tests seemed to pass for me.
P.S. I hope I'm following the right procedure, I've never contributed before so please let me know if I'm missing anything.