This repository has been archived by the owner on Jan 20, 2023. It is now read-only.
Fix logic when adding a value which could go into multiple centroids #7
Conversation
This file contains 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
When adding a value, we pick a centroid for it to be added to. Usually, there is just one option, the centroid with the nearest mean. In some cases (typically when a TDigest has very few values), it's possible for multiple centroids to have means which are equidistant from the input value. The previous logic for this was broken in that it did not carefully maintain the invariant that the centroid means should always be non-decreasing. Previously, we would first filter the candidate set to only a set of centroid which have room for more values, and then we would filter again to preserve the invariant. This was backwards. Instead, we need to first make sure the invariant is preserved. If the right centroid is full, we just need to add a new one. Other logic (in addNewCentroid) will make sure it's added in the right spot.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
When adding a value, we pick a centroid for it to be added to. Usually, there is just one option, the centroid with the nearest mean. In some cases (typically when a TDigest has very few values), it's possible for multiple centroids to have means which are equidistant from the input value.
The previous logic for this was broken in that it did not carefully maintain the invariant that the centroid means should always be non-decreasing. Previously, we would first filter the candidate set to only a set of centroid which have room for more values, and then we would filter again to preserve the invariant. This was backwards.
Instead, we need to first make sure the invariant is preserved. If the right centroid is full, we just need to add a new one. Other logic (in
addNewCentroid) will make sure it's added in the right spot.Fixes #6.