Merge branch 'master' of github.com:/RubixML/ML

andrewdalpino · andrewdalpino · commit 1f3bdec7c154 · 2023-07-05T19:30:09.000-05:00
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,6 @@
 - 2.4.1
     - Sentence Tokenizer fix Arabic and Farsi language support
+    - Optimize online variance updating
 
 - 2.4.0
     - Add GELU activation function
diff --git a/benchmarks/AnomalyDetectors/GaussianMLEBench.php b/benchmarks/AnomalyDetectors/GaussianMLEBench.php
@@ -12,7 +12,7 @@
  */
 class GaussianMLEBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/AnomalyDetectors/IsolationForestBench.php b/benchmarks/AnomalyDetectors/IsolationForestBench.php
@@ -12,7 +12,7 @@
  */
 class IsolationForestBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/AnomalyDetectors/LocalOutlierFactorBench.php b/benchmarks/AnomalyDetectors/LocalOutlierFactorBench.php
@@ -12,7 +12,7 @@
  */
 class LocalOutlierFactorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/AnomalyDetectors/LodaBench.php b/benchmarks/AnomalyDetectors/LodaBench.php
@@ -12,7 +12,7 @@
  */
 class LodaBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/AnomalyDetectors/OneClassSVMBench.php b/benchmarks/AnomalyDetectors/OneClassSVMBench.php
@@ -12,7 +12,7 @@
  */
 class OneClassSVMBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/AnomalyDetectors/RobustZScoreBench.php b/benchmarks/AnomalyDetectors/RobustZScoreBench.php
@@ -12,7 +12,7 @@
  */
 class RobustZScoreBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/AdaBoostBench.php b/benchmarks/Classifiers/AdaBoostBench.php
@@ -13,7 +13,7 @@
  */
 class AdaBoostBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/ClassificationTreeBench.php b/benchmarks/Classifiers/ClassificationTreeBench.php
@@ -12,7 +12,7 @@
  */
 class ClassificationTreeBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/ExtraTreeClassifierBench.php b/benchmarks/Classifiers/ExtraTreeClassifierBench.php
@@ -12,7 +12,7 @@
  */
 class ExtraTreeClassifierBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/GaussianNBBench.php b/benchmarks/Classifiers/GaussianNBBench.php
@@ -12,7 +12,7 @@
  */
 class GaussianNBBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/KDNeighborsBench.php b/benchmarks/Classifiers/KDNeighborsBench.php
@@ -12,7 +12,7 @@
  */
 class KDNeighborsBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/KNearestNeighborsBench.php b/benchmarks/Classifiers/KNearestNeighborsBench.php
@@ -12,7 +12,7 @@
  */
 class KNearestNeighborsBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/LogisticRegressionBench.php b/benchmarks/Classifiers/LogisticRegressionBench.php
@@ -12,7 +12,7 @@
  */
 class LogisticRegressionBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/LogitBoostBench.php b/benchmarks/Classifiers/LogitBoostBench.php
@@ -12,7 +12,7 @@
  */
 class LogitBoostBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/MultilayerPerceptronBench.php b/benchmarks/Classifiers/MultilayerPerceptronBench.php
@@ -15,7 +15,7 @@
  */
 class MultilayerPerceptronBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/NaiveBayesBench.php b/benchmarks/Classifiers/NaiveBayesBench.php
@@ -13,7 +13,7 @@
  */
 class NaiveBayesBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/OneVsRestBench.php b/benchmarks/Classifiers/OneVsRestBench.php
@@ -14,7 +14,7 @@
  */
 class OneVsRestBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/RadiusNeighborsBench.php b/benchmarks/Classifiers/RadiusNeighborsBench.php
@@ -12,7 +12,7 @@
  */
 class RadiusNeighborsBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/RandomForestBench.php b/benchmarks/Classifiers/RandomForestBench.php
@@ -13,7 +13,7 @@
  */
 class RandomForestBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/SVCBench.php b/benchmarks/Classifiers/SVCBench.php
@@ -13,7 +13,7 @@
  */
 class SVCBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Classifiers/SoftmaxClassifierBench.php b/benchmarks/Classifiers/SoftmaxClassifierBench.php
@@ -12,7 +12,7 @@
  */
 class SoftmaxClassifierBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Clusterers/FuzzyCMeansBench.php b/benchmarks/Clusterers/FuzzyCMeansBench.php
@@ -12,7 +12,7 @@
  */
 class FuzzyCMeansBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Clusterers/GaussianMixtureBench.php b/benchmarks/Clusterers/GaussianMixtureBench.php
@@ -12,7 +12,7 @@
  */
 class GaussianMixtureBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Clusterers/KMeansBench.php b/benchmarks/Clusterers/KMeansBench.php
@@ -12,7 +12,7 @@
  */
 class KMeansBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Clusterers/MeanShiftBench.php b/benchmarks/Clusterers/MeanShiftBench.php
@@ -12,7 +12,7 @@
  */
 class MeanShiftBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Persisters/Serializers/GzipNativeBench.php b/benchmarks/Persisters/Serializers/GzipNativeBench.php
@@ -13,7 +13,7 @@
  */
 class GzipNativeBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     /**
      * @var \Rubix\ML\Serializers\GzipNative
diff --git a/benchmarks/Persisters/Serializers/NativeBench.php b/benchmarks/Persisters/Serializers/NativeBench.php
@@ -13,7 +13,7 @@
  */
 class NativeBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     /**
      * @var \Rubix\ML\Serializers\Native
diff --git a/benchmarks/Persisters/Serializers/RBXBench.php b/benchmarks/Persisters/Serializers/RBXBench.php
@@ -13,7 +13,7 @@
  */
 class RBXBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     /**
      * @var \Rubix\ML\Serializers\RBX
diff --git a/benchmarks/Regressors/AdalineBench.php b/benchmarks/Regressors/AdalineBench.php
@@ -11,7 +11,7 @@
  */
 class AdalineBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/ExtraTreeRegressorBench.php b/benchmarks/Regressors/ExtraTreeRegressorBench.php
@@ -11,7 +11,7 @@
  */
 class ExtraTreeRegressorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/GradientBoostBench.php b/benchmarks/Regressors/GradientBoostBench.php
@@ -11,7 +11,7 @@
  */
 class GradientBoostBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/KDNeighborsRegressorBench.php b/benchmarks/Regressors/KDNeighborsRegressorBench.php
@@ -11,7 +11,7 @@
  */
 class KDNeighborsRegressorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/KNNRegressorBench.php b/benchmarks/Regressors/KNNRegressorBench.php
@@ -11,7 +11,7 @@
  */
 class KNNRegressorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/MLPRegressorBench.php b/benchmarks/Regressors/MLPRegressorBench.php
@@ -14,7 +14,7 @@
  */
 class MLPRegressorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/RadiusNeighborsRegressorBench.php b/benchmarks/Regressors/RadiusNeighborsRegressorBench.php
@@ -11,7 +11,7 @@
  */
 class RadiusNeighborsRegressorBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/RegressionTreeBench.php b/benchmarks/Regressors/RegressionTreeBench.php
@@ -11,7 +11,7 @@
  */
 class RegressionTreeBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/RidgeBench.php b/benchmarks/Regressors/RidgeBench.php
@@ -11,7 +11,7 @@
  */
 class RidgeBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/benchmarks/Regressors/SVRBench.php b/benchmarks/Regressors/SVRBench.php
@@ -11,7 +11,7 @@
  */
 class SVRBench
 {
-    protected const TRAINING_SIZE = 2500;
+    protected const TRAINING_SIZE = 10000;
 
     protected const TESTING_SIZE = 10000;
 
diff --git a/src/AnomalyDetectors/GaussianMLE.php b/src/AnomalyDetectors/GaussianMLE.php
@@ -244,6 +244,8 @@ public function partial(Dataset $dataset) : void
 
         $n = $dataset->numSamples();
 
+        $weight = $this->n + $n;
+
         foreach ($dataset->features() as $column => $values) {
             [$mean, $variance] = Stats::meanVar($values);
 
@@ -253,13 +255,13 @@ public function partial(Dataset $dataset) : void
             $oldVariance -= $this->epsilon;
 
             $this->means[$column] = (($this->n * $oldMean)
-                + ($n * $mean)) / ($this->n + $n);
+                + ($n * $mean)) / $weight;
 
             $this->variances[$column] = ($this->n
                 * $oldVariance + ($n * $variance)
-                + ($this->n / ($n * ($this->n + $n)))
+                + ($this->n / ($n * $weight))
                 * ($n * $oldMean - $n * $mean) ** 2)
-                / ($this->n + $n);
+                / $weight;
         }
 
         $epsilon = max($this->smoothing * max($this->variances), CPU::epsilon());
@@ -270,15 +272,15 @@ public function partial(Dataset $dataset) : void
 
         $this->epsilon = $epsilon;
 
-        $this->n += $n;
+        $this->n = $weight;
 
         $lls = array_map([$this, 'logLikelihood'], $dataset->samples());
 
         $threshold = Stats::quantile($lls, 1.0 - $this->contamination);
 
-        $weight = $n / $this->n;
+        $proportion = $n / $this->n;
 
-        $this->threshold = (1.0 - $weight) * $this->threshold + $weight * $threshold;
+        $this->threshold = $proportion * $threshold + (1.0 - $proportion) * $this->threshold;
     }
 
     /**
diff --git a/src/Classifiers/GaussianNB.php b/src/Classifiers/GaussianNB.php
diff --git a/src/Transformers/ZScaleStandardizer.php b/src/Transformers/ZScaleStandardizer.php

Original file line number	Diff line number	Diff line change
`@@ -12,7 +12,7 @@`
`12`	`12`	`*/`
`13`	`13`	`class GaussianMLEBench`
`14`	`14`	`{`
`15`		`- protected const TRAINING_SIZE = 2500;`
	`15`	`+ protected const TRAINING_SIZE = 10000;`
`16`	`16`
`17`	`17`	`protected const TESTING_SIZE = 10000;`
`18`	`18`
Original file line number	Diff line number	Diff line change
`@@ -13,7 +13,7 @@`
`13`	`13`	`*/`
`14`	`14`	`class AdaBoostBench`
`15`	`15`	`{`
`16`		`- protected const TRAINING_SIZE = 2500;`
	`16`	`+ protected const TRAINING_SIZE = 10000;`
`17`	`17`
`18`	`18`	`protected const TESTING_SIZE = 10000;`
`19`	`19`
Original file line number	Diff line number	Diff line change
`@@ -15,7 +15,7 @@`
`15`	`15`	`*/`
`16`	`16`	`class MultilayerPerceptronBench`
`17`	`17`	`{`
`18`		`- protected const TRAINING_SIZE = 2500;`
	`18`	`+ protected const TRAINING_SIZE = 10000;`
`19`	`19`
`20`	`20`	`protected const TESTING_SIZE = 10000;`
`21`	`21`
Original file line number	Diff line number	Diff line change
`@@ -14,7 +14,7 @@`
`14`	`14`	`*/`
`15`	`15`	`class OneVsRestBench`
`16`	`16`	`{`
`17`		`- protected const TRAINING_SIZE = 2500;`
	`17`	`+ protected const TRAINING_SIZE = 10000;`
`18`	`18`
`19`	`19`	`protected const TESTING_SIZE = 10000;`
`20`	`20`
Original file line number	Diff line number	Diff line change
`@@ -11,7 +11,7 @@`
`11`	`11`	`*/`
`12`	`12`	`class AdalineBench`
`13`	`13`	`{`
`14`		`- protected const TRAINING_SIZE = 2500;`
	`14`	`+ protected const TRAINING_SIZE = 10000;`
`15`	`15`
`16`	`16`	`protected const TESTING_SIZE = 10000;`
`17`	`17`