Forego unnecessary logistic computation in Logit Boost

Author: andrewdalpino

CHANGELOG.md

@@ -1,3 +1,6 @@
+- 1.3.3
+    - Forego unnecessary logistic computation in Logit Boost
+
 - 1.3.2
     - Optimize Binary output layer
 

src/Classifiers/ExtraTreeClassifier.php

@@ -213,8 +213,7 @@ public function proba(Dataset $dataset) : array
     }
 
     /**
-     * Terminate the branch by selecting the class outcome with the highest
-     * probability.
+     * Terminate the branch by selecting the class outcome with the highest probability.
      *
      * @param \Rubix\ML\Datasets\Labeled $dataset
      * @return \Rubix\ML\Graph\Nodes\Best
@@ -235,9 +234,9 @@ protected function terminate(Labeled $dataset) : Best
 
         $p = $counts[$outcome] / $n;
 
-        $impurity = -($p * log($p));
+        $entropy = -($p * log($p));
 
-        return new Best($outcome, $probabilities, $impurity, $n);
+        return new Best($outcome, $probabilities, $entropy, $n);
     }
 
     /**

src/Classifiers/LogisticRegression.php

@@ -57,7 +57,7 @@ class LogisticRegression implements Estimator, Learner, Online, Probabilistic, R
     /**
      * The number of training samples to process at a time.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

src/Classifiers/LogitBoost.php

@@ -427,7 +427,6 @@ public function train(Dataset $dataset) : void
             $zHat = $booster->predict($training);
 
             $z = array_map([$this, 'updateZ'], $zHat, $z);
-
             $out = array_map('Rubix\ML\sigmoid', $z);
 
             $this->losses[$epoch] = $loss;
@@ -439,12 +438,10 @@ public function train(Dataset $dataset) : void
 
                 $zTest = array_map([$this, 'updateZ'], $zHat, $zTest);
 
-                $outTest = array_map('Rubix\ML\sigmoid', $zTest);
-
                 $predictions = [];
 
-                foreach ($outTest as $probability) {
-                    $predictions[] = $probability < 0.5 ? $classes[0] : $classes[1];
+                foreach ($zTest as $value) {
+                    $predictions[] = $value < 0.0 ? $classes[0] : $classes[1];
                 }
 
                 $score = $this->metric->score($predictions, $testing->labels());

src/Classifiers/MultilayerPerceptron.php

@@ -74,7 +74,7 @@ class MultilayerPerceptron implements Estimator, Learner, Online, Probabilistic,
     /**
      * The number of training samples to process at a time.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

src/Classifiers/SoftmaxClassifier.php

@@ -54,7 +54,7 @@ class SoftmaxClassifier implements Estimator, Learner, Online, Probabilistic, Ve
     /**
      * The number of training samples to process at a time.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

src/Clusterers/KMeans.php

@@ -66,7 +66,7 @@ class KMeans implements Estimator, Learner, Online, Probabilistic, Verbose, Pers
     /**
      * The size of each mini batch in samples.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

src/Datasets/Labeled.php

@@ -570,7 +570,7 @@ public function stratifiedFold(int $k = 10) : array
      * not enough samples to fill an entire batch, then the dataset will contain
      * as many samples and labels as possible.
      *
-     * @param int $n
+     * @param positive-int $n
      * @return list<self>
      */
     public function batch(int $n = 50) : array
@@ -742,7 +742,8 @@ public function randomWeightedSubsetWithReplacement(int $n, array $weights) : se
                 . ' but ' . count($weights) . ' given.');
         }
 
-        $numLevels = (int) round(sqrt(count($weights)));
+        /** @var positive-int $numLevels */
+        $numLevels = (int) round(sqrt(count($weights))) ?: 1;
 
         $levels = array_chunk($weights, $numLevels, true);
 

src/Datasets/Unlabeled.php

@@ -303,7 +303,7 @@ public function fold(int $k = 3) : array
      * not enough samples to fill an entire batch, then the dataset will contain
      * as many samples as possible.
      *
-     * @param int $n
+     * @param positive-int $n
      * @return list<self>
      */
     public function batch(int $n = 50) : array
@@ -453,6 +453,7 @@ public function randomWeightedSubsetWithReplacement(int $n, array $weights) : se
                 . ' but ' . count($weights) . ' given.');
         }
 
+        /** @var positive-int $numLevels */
         $numLevels = (int) round(sqrt(count($weights)));
 
         $levels = array_chunk($weights, $numLevels, true);

src/Regressors/Adaline.php

@@ -58,7 +58,7 @@ class Adaline implements Estimator, Learner, Online, RanksFeatures, Verbose, Per
     /**
      * The number of training samples to process at a time.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

src/Regressors/MLPRegressor.php

@@ -73,7 +73,7 @@ class MLPRegressor implements Estimator, Learner, Online, Verbose, Persistable
     /**
      * The number of training samples to process at a time.
      *
-     * @var int
+     * @var positive-int
      */
     protected int $batchSize;
 

tests/Classifiers/LogitBoostTest.php

@@ -76,8 +76,8 @@ class LogitBoostTest extends TestCase
     protected function setUp() : void
     {
         $this->generator = new Agglomerate([
-            'male' => new Blob([69.2, 195.7, 40.0], [2.0, 6.0, 0.6]),
-            'female' => new Blob([63.7, 168.5, 38.1], [1.6, 5.0, 0.8]),
+            'male' => new Blob([69.2, 195.7, 40.0], [2.0, 6.4, 0.6]),
+            'female' => new Blob([63.7, 168.5, 38.1], [1.8, 5.0, 0.8]),
         ]);
 
         $this->estimator = new LogitBoost(new RegressionTree(3), 0.1, 0.5, 1000, 1e-4, 5, 0.1, new FBeta());