Skip to content

Add multiclassification topk and confmatrix metrics to model insights serialization format #537

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Jan 15, 2021
Merged

Add multiclassification topk and confmatrix metrics to model insights serialization format #537

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
7c4b3e0
add topk and confusion matrix metrics to model insights serialization…
feifjiang Jan 5, 2021
abc240d
modified threshold for flaky RecordInisghtsLOCOTest
feifjiang Jan 5, 2021
e037c3f
(1)add case class ClassCount to represent MisClassifications (2)add t…
feifjiang Jan 8, 2021
33184bf
Merge branch 'master' into ff/insights
tovbinm Jan 8, 2021
9ea7e44
remove unused library imports
feifjiang Jan 8, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion core/src/main/scala/com/salesforce/op/ModelInsights.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -401,7 +401,9 @@ case object ModelInsights {
classOf[DataBalancerSummary], classOf[DataCutterSummary], classOf[DataSplitterSummary],
classOf[SingleMetric], classOf[MultiMetrics], classOf[BinaryClassificationMetrics],
classOf[BinaryClassificationBinMetrics], classOf[MulticlassThresholdMetrics],
classOf[BinaryThresholdMetrics], classOf[MultiClassificationMetrics], classOf[RegressionMetrics]
classOf[BinaryThresholdMetrics], classOf[MultiClassificationMetrics], classOf[RegressionMetrics],
classOf[MultiClassificationMetricsTopK],
classOf[MulticlassConfMatrixMetricsByThreshold], classOf[MisClassificationMetrics]
))
val evalMetricsSerializer = new CustomSerializer[EvalMetric](_ =>
( { case JString(s) => EvalMetric.withNameInsensitive(s) },
Expand Down
4 changes: 2 additions & 2 deletions core/src/test/scala/com/salesforce/op/stages/impl/insights/RecordInsightsLOCOTest.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -232,8 +232,8 @@ class RecordInsightsLOCOTest extends FunSpec with TestSparkContext with RecordIn
info("Each feature vector should only have either three or four non-zero entries. One each from country and " +
"picklist, while currency can have either two (if it's null the currency column will be filled with the mean)" +
" or just one if it's not null.")
it("should pick between 1 and 4 of the features") {
all(parsed.map(_.size)) should (be >= 1 and be <= 4)
it("should pick between 0 and 4 of the features") {
Copy link
Collaborator

@tovbinm tovbinm Jan 5, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why did this change?

Copy link
Contributor Author

@feifjiang feifjiang Jan 8, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tovbinm this test is quite flaky. In one of the failed runs, the value was 0 and not between 1 and 4.

Copy link
Collaborator

@tovbinm tovbinm Jan 8, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@leahmcguire @Jauntbox do you have any ideas?

Copy link
Contributor

@nicodv nicodv Jan 13, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There's still an open PR to address this: #423

We could use a better test, but for now we can start by getting rid of the flakiness.

all(parsed.map(_.size)) should (be >= 0 and be <= 4)
}

// Grab the feature vector metadata for comparison against the LOCO record insights
Expand Down