Confidence Intervals for Permutation Feature Importance #1844
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
added 2 commits
December 6, 2018 14:01
|
@glebuk @TomFinley I'd appreciate a look at this if possible. I added a new class that returns descriptive statistics. What do you two think? |
glebuk
reviewed
Dec 7, 2018
docs/samples/Microsoft.ML.Samples/Dynamic/PermutationFeatureImportance.cs
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
…tics package. Updated the sample to use the standard error.
rogancarr
changed the title
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 7, 2018
Ivanidzo4ka
reviewed
Dec 11, 2018
| public abstract class MetricsStatisticsBase<T>{ | ||
| internal MetricsStatisticsBase() | ||
| { | ||
| } |
There was a problem hiding this comment.
why you need this constructor?
There was a problem hiding this comment.
I have to make this class public, but I don't want it to be inheritable outside this namespace. Is there a better way to restrict it?
Ivanidzo4ka
reviewed
Dec 11, 2018
docs/samples/Microsoft.ML.Samples/Dynamic/PermutationFeatureImportance.cs
Outdated
Show resolved
Hide resolved
docs/samples/Microsoft.ML.Samples/Dynamic/PermutationFeatureImportance.cs
Outdated
Show resolved
Hide resolved
glebuk
reviewed
Dec 18, 2018
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Outdated
Show resolved
Hide resolved
justinormont
approved these changes
Dec 20, 2018
ghost
locked as resolved and limited conversation to collaborators
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.
This PR adds an optional confidence interval calculation to the
Permutation Feature Importanceevaluator. If the users specifies the number of permutations > 1, then the resulting evaluating metrics will contain the mean and standard deviation.Fixes #1840
Update: In this PR, I am removing PFI for clustering. It actually doesn't make sense to use clustering metrics with PFI because cluster membership can change during permutations. If we want to look at feature importance for clustering, it'll be better to use a test specific to clustering.