Updated EmpiricalProbabilityCalibrationPoint to be named EmpiricalAndPredictedProbability.

src/Learners/Classifier/ClassifierEvaluator.cs

@@ -775,7 +775,7 @@ namespace Microsoft.ML.Probabilistic.Learners
         /// <param name="instanceSource">The instance source.</param>
         /// <param name="predictions">The predictions.</param>
         /// <returns>The computed empirical probability calibration curve.</returns>
-        public IEnumerable<EmpiricalProbabilityCalibrationPoint> CalibrationCurve(
+        public IEnumerable<EmpiricalAndPredictedProbability> CalibrationCurve(
             TLabel positiveClassLabel,
             TInstanceSource instanceSource,
             IEnumerable<IDictionary<TLabel, double>> predictions)
@@ -792,7 +792,7 @@ namespace Microsoft.ML.Probabilistic.Learners
         /// <param name="binCount">The number of bins to use.</param>
         /// <param name="minBinInstanceCount">The minimal number of instances per bin. Defaults to 1.</param>
         /// <returns>The computed empirical probability calibration curve.</returns>
-        public IEnumerable<EmpiricalProbabilityCalibrationPoint> CalibrationCurve(
+        public IEnumerable<EmpiricalAndPredictedProbability> CalibrationCurve(
             TLabel positiveClassLabel,
             TInstanceSource instanceSource,
             IEnumerable<IDictionary<TLabel, double>> predictions,
@@ -810,7 +810,7 @@ namespace Microsoft.ML.Probabilistic.Learners
         /// <param name="labelSource">The label source.</param>
         /// <param name="predictions">The predictions.</param>
         /// <returns>The computed empirical probability calibration curve.</returns>
-        public IEnumerable<EmpiricalProbabilityCalibrationPoint> CalibrationCurve(
+        public IEnumerable<EmpiricalAndPredictedProbability> CalibrationCurve(
             TLabel positiveClassLabel,
             TInstanceSource instanceSource,
             TLabelSource labelSource,
@@ -843,7 +843,7 @@ namespace Microsoft.ML.Probabilistic.Learners
         /// <param name="binCount">The number of bins to use.</param>
         /// <param name="minBinInstanceCount">The minimal number of instances per bin. Defaults to 1.</param>
         /// <returns>The computed empirical probability calibration curve.</returns>
-        public IEnumerable<EmpiricalProbabilityCalibrationPoint> CalibrationCurve(
+        public IEnumerable<EmpiricalAndPredictedProbability> CalibrationCurve(
             TLabel positiveClassLabel, 
             TInstanceSource instanceSource,
             TLabelSource labelSource,
@@ -916,7 +916,7 @@ namespace Microsoft.ML.Probabilistic.Learners
         /// <param name="groundTruthInstances">The ground truth instances.</param>
         /// <param name="minBinInstanceCount">The minimal number of instances per bin. Defaults to 1.</param>
         /// <returns>The computed empirical probability calibration curve.</returns>
-        private IEnumerable<EmpiricalProbabilityCalibrationPoint> CalibrationCurve(
+        private IEnumerable<EmpiricalAndPredictedProbability> CalibrationCurve(
             TLabel positiveClassLabel,
             TInstanceSource instanceSource,
             TLabelSource labelSource,
@@ -957,12 +957,12 @@ namespace Microsoft.ML.Probabilistic.Learners
                 }
             }
 
-            var calibrationCurve = new List<EmpiricalProbabilityCalibrationPoint>();
+            var calibrationCurve = new List<EmpiricalAndPredictedProbability>();
             for (int bin = 0; bin < binCount; bin++)
             {
                 if (predictedCount[bin] >= minBinInstanceCount)
                 {
-                    calibrationCurve.Add(new EmpiricalProbabilityCalibrationPoint(
+                    calibrationCurve.Add(new EmpiricalAndPredictedProbability(
                         (bin + 0.5) / binCount,
                         (double)positiveClassCount[bin] / predictedCount[bin]));
                 }

src/Runtime/Core/Collections/EmpiricalAndPredictedProbability.cs

@@ -6,14 +6,14 @@ namespace Microsoft.ML.Probabilistic.Collections
 {
     using Microsoft.ML.Probabilistic.Utilities;
 
-    public struct EmpiricalProbabilityCalibrationPoint
+    public struct EmpiricalAndPredictedProbability
     {
         /// <summary>
-        /// Initializes a new instance of the <see cref="EmpiricalProbabilityCalibrationPoint"/> struct.
+        /// Initializes a new instance of the <see cref="EmpiricalAndPredictedProbability"/> struct.
         /// </summary>
         /// <param name="empiricalProbability">The empirical probability</param>
         /// <param name="predictedProbability">The predicted probability</param>
-        public EmpiricalProbabilityCalibrationPoint(double empiricalProbability, double predictedProbability)
+        public EmpiricalAndPredictedProbability(double empiricalProbability, double predictedProbability)
             : this()
         {
             this.EmpiricalProbability = empiricalProbability;
@@ -31,9 +31,9 @@ namespace Microsoft.ML.Probabilistic.Collections
         public readonly double PredictedProbability;
         
         /// <summary>
-        /// Gets the string representation of this <see cref="EmpiricalProbabilityCalibrationPoint"/>.
+        /// Gets the string representation of this <see cref="EmpiricalAndPredictedProbability"/>.
         /// </summary>
-        /// <returns>The string representation of the <see cref="EmpiricalProbabilityCalibrationPoint"/>.</returns>
+        /// <returns>The string representation of the <see cref="EmpiricalAndPredictedProbability"/>.</returns>
         public override string ToString()
         {
             return $"{this.EmpiricalProbability}, {this.PredictedProbability}";
@@ -49,7 +49,7 @@ namespace Microsoft.ML.Probabilistic.Collections
         /// </returns>
         public override bool Equals(object obj)
         {
-            if (obj is EmpiricalProbabilityCalibrationPoint calibrationPoint)
+            if (obj is EmpiricalAndPredictedProbability calibrationPoint)
             {
                 return object.Equals(this.EmpiricalProbability, calibrationPoint.EmpiricalProbability) && object.Equals(this.PredictedProbability, calibrationPoint.PredictedProbability);
             }

test/Learners/LearnersTests/ClassifierEvaluatorTests.cs

@@ -252,18 +252,18 @@ namespace Microsoft.ML.Probabilistic.Learners.Tests
         public void CalibrationCurveTest()
         {
             // Curve for perfect predictions
-            var expected = new[] { new EmpiricalProbabilityCalibrationPoint(0.25, 0.0), new EmpiricalProbabilityCalibrationPoint(0.75, 1.0) };
+            var expected = new[] { new EmpiricalAndPredictedProbability(0.25, 0.0), new EmpiricalAndPredictedProbability(0.75, 1.0) };
             var actual = this.evaluator.CalibrationCurve(LabelSet[0], this.groundTruth, this.groundTruth).ToArray();
             Xunit.Assert.Equal(expected, actual);
 
             // Curve for imperfect predictions (one-versus-rest)
-            expected = new[] { new EmpiricalProbabilityCalibrationPoint(0.25, 0.75), new EmpiricalProbabilityCalibrationPoint(0.75, 0.0) };
+            expected = new[] { new EmpiricalAndPredictedProbability(0.25, 0.75), new EmpiricalAndPredictedProbability(0.75, 0.0) };
             actual = this.evaluator.CalibrationCurve(LabelSet[0], this.groundTruth, this.predictions).ToArray();
             Xunit.Assert.Equal(expected, actual);
 
             // Curve for imperfect predictions (3 bins)
             const int BinCount = 4;
-            expected = new[] { new EmpiricalProbabilityCalibrationPoint(1 / 8.0, 0.75), new EmpiricalProbabilityCalibrationPoint(7 / 8.0, 0.0) };
+            expected = new[] { new EmpiricalAndPredictedProbability(1 / 8.0, 0.75), new EmpiricalAndPredictedProbability(7 / 8.0, 0.0) };
             actual = this.evaluator.CalibrationCurve(LabelSet[0], this.groundTruth, this.predictions, BinCount).ToArray();
             Xunit.Assert.Equal(expected, actual);