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Outdated suggestions cannot be applied.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.
I am running fits, hoping it converges.
Haven't looked at results yet.
In the following example's current form, it is too slow for CI:
Weights are currently constrained to sum to 1, but they are also allowed to be negative, e.g. if you have two models, you could end up with a solution like
2 .* sol1 .- sol2
.We could place it on a simplex instead, but I figured to start with, it'd be interesting to see if it does like such solutions where you subtract a result.
My concern is that this could be too much freedom/allow for creative overfitting. I won'd want to see solutions like