New metric: Precision at fixed recall

CHANGELOG.md

@@ -88,6 +88,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added `ModifiedPanopticQuality` metric to detection package ([#1627](https://github.com/Lightning-AI/metrics/pull/1627))
 
 
+- Added `PrecisionAtFixedRecall` metric to classification package ([#1683](https://github.com/Lightning-AI/torchmetrics/pull/1683))
+
+
 ### Changed
 
 - Changed `update_count` and `update_called` from private to public methods ([#1370](https://github.com/Lightning-AI/metrics/pull/1370))

docs/source/classification/precision_at_fixed_recall.rst

@@ -0,0 +1,53 @@
+.. customcarditem::
+   :header: Recall At Fixed Precision
+   :image: https://pl-flash-data.s3.amazonaws.com/assets/thumbnails/tabular_classification.svg
+   :tags: Classification
+
+#########################
+Precision At Fixed Recall
+#########################
+
+Module Interface
+________________
+
+BinaryPrecisionAtFixedRecall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autoclass:: torchmetrics.classification.BinaryPrecisionAtFixedRecall
+    :noindex:
+    :exclude-members: update, compute
+
+MulticlassPrecisionAtFixedRecall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autoclass:: torchmetrics.classification.MulticlassPrecisionAtFixedRecall
+    :noindex:
+    :exclude-members: update, compute
+
+MultilabelPrecisionAtFixedRecall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autoclass:: torchmetrics.classification.MultilabelPrecisionAtFixedRecall
+    :noindex:
+    :exclude-members: update, compute
+
+Functional Interface
+____________________
+
+binary_precision_at_fixed_recall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autofunction:: torchmetrics.functional.classification.binary_precision_at_fixed_recall
+    :noindex:
+
+multiclass_precision_at_fixed_recall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autofunction:: torchmetrics.functional.classification.multiclass_precision_at_fixed_recall
+    :noindex:
+
+multilabel_precision_at_fixed_recall
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autofunction:: torchmetrics.functional.classification.multilabel_precision_at_fixed_recall
+    :noindex:

src/torchmetrics/__init__.py

@@ -39,8 +39,10 @@
     JaccardIndex,
     MatthewsCorrCoef,
     Precision,
+    PrecisionAtFixedRecall,
     PrecisionRecallCurve,
     Recall,
+    RecallAtFixedPrecision,
     Specificity,
     StatScores,
 )
@@ -173,10 +175,12 @@
     "PermutationInvariantTraining",
     "Perplexity",
     "Precision",
+    "PrecisionAtFixedRecall",
     "PrecisionRecallCurve",
     "PeakSignalNoiseRatio",
     "R2Score",
     "Recall",
+    "RecallAtFixedPrecision",
     "RelativeAverageSpectralError",
     "RetrievalFallOut",
     "RetrievalHitRate",

src/torchmetrics/classification/__init__.py

@@ -75,6 +75,12 @@
     MulticlassMatthewsCorrCoef,
     MultilabelMatthewsCorrCoef,
 )
+from torchmetrics.classification.precision_fixed_recall import (
+    BinaryPrecisionAtFixedRecall,
+    MulticlassPrecisionAtFixedRecall,
+    MultilabelPrecisionAtFixedRecall,
+    PrecisionAtFixedRecall,
+)
 from torchmetrics.classification.precision_recall import (
     BinaryPrecision,
     BinaryRecall,
@@ -90,10 +96,11 @@
     MultilabelRankingAveragePrecision,
     MultilabelRankingLoss,
 )
-from torchmetrics.classification.recall_at_fixed_precision import (
+from torchmetrics.classification.recall_fixed_precision import (
     BinaryRecallAtFixedPrecision,
     MulticlassRecallAtFixedPrecision,
     MultilabelRecallAtFixedPrecision,
+    RecallAtFixedPrecision,
 )
 from torchmetrics.classification.roc import ROC, BinaryROC, MulticlassROC, MultilabelROC
 from torchmetrics.classification.specificity import (
@@ -102,7 +109,7 @@
     MultilabelSpecificity,
     Specificity,
 )
-from torchmetrics.classification.specificity_at_sensitivity import (
+from torchmetrics.classification.specificity_sensitivity import (
     BinarySpecificityAtSensitivity,
     MulticlassSpecificityAtSensitivity,
     MultilabelSpecificityAtSensitivity,
@@ -199,4 +206,12 @@
     "BinarySpecificityAtSensitivity",
     "MulticlassSpecificityAtSensitivity",
     "MultilabelSpecificityAtSensitivity",
+    "BinaryPrecisionAtFixedRecall",
+    "MulticlassPrecisionAtFixedRecall",
+    "MultilabelPrecisionAtFixedRecall",
+    "PrecisionAtFixedRecall",
+    "RecallAtFixedPrecision",
+    "BinaryPrecisionAtFixedRecall",
+    "MulticlassPrecisionAtFixedRecall",
+    "MultilabelPrecisionAtFixedRecall",
 ]