Fix Naive Bayes divide by zero when smoothing is 0

CHANGELOG.md

@@ -1,6 +1,7 @@
 - 0.1.5
     - Compensate for zero vectors in Cosine kernel
     - Fixed KMC2 random threshold calculation
+    - Fix Naive Bayes divide by zero when smoothing is 0
 
 - 0.1.4
     - Optimized Cosine distance for sparse vectors

docs/classifiers/naive-bayes.md

@@ -12,7 +12,7 @@ Categorical Naive Bayes is a probability-based classifier that uses counting and
 ## Parameters
 | # | Param | Default | Type | Description |
 |---|---|---|---|---|
-| 1 | alpha | 1.0 | float | The amount of (Laplace) smoothing added to the probabilities. |
+| 1 | smoothing | 1.0 | float | The amount of (Laplace) smoothing added to the probabilities. |
 | 2 | priors | null | array | The class prior probabilities as an associative array with class labels as keys and the prior probabilities as values. If null, then the learner will compute these values from the training data. |
 
 ## Example

src/Classifiers/NaiveBayes.php

@@ -51,7 +51,7 @@ class NaiveBayes implements Estimator, Learner, Online, Probabilistic, Persistab
      *
      * @var float
      */
-    protected $alpha;
+    protected $smoothing;
 
     /**
      * The class prior log probabilities.
@@ -97,15 +97,15 @@ class NaiveBayes implements Estimator, Learner, Online, Probabilistic, Persistab
     ];
 
     /**
-     * @param float $alpha
+     * @param float $smoothing
      * @param (int|float)[]|null $priors
      * @throws \InvalidArgumentException
      */
-    public function __construct(float $alpha = 1.0, ?array $priors = null)
+    public function __construct(float $smoothing = 1.0, ?array $priors = null)
     {
-        if ($alpha < 0.0) {
-            throw new InvalidArgumentException('Alpha must be'
-                . " greater than 0, $alpha given.");
+        if ($smoothing <= 0.0) {
+            throw new InvalidArgumentException('Smoothing must be'
+                . " greater than 0, $smoothing given.");
         }
 
         $logPriors = [];
@@ -128,7 +128,7 @@ public function __construct(float $alpha = 1.0, ?array $priors = null)
             }
         }
 
-        $this->alpha = $alpha;
+        $this->smoothing = $smoothing;
         $this->logPriors = $logPriors;
         $this->fitPriors = is_null($priors);
     }
@@ -163,7 +163,7 @@ public function compatibility() : array
     public function params() : array
     {
         return [
-            'alpha' => $this->alpha,
+            'smoothing' => $this->smoothing,
             'priors' => $this->fitPriors ? null : $this->priors(),
         ];
     }
@@ -250,12 +250,12 @@ public function partial(Dataset $dataset) : void
                     }
                 }
 
-                $total = array_sum($columnCounts) + (count($columnCounts) * $this->alpha);
+                $total = array_sum($columnCounts) + (count($columnCounts) * $this->smoothing);
 
                 $probs = [];
 
                 foreach ($columnCounts as $category => $count) {
-                    $probs[$category] = log(($count + $this->alpha) / $total);
+                    $probs[$category] = log(($count + $this->smoothing) / $total);
                 }
 
                 $classCounts[$column] = $columnCounts;
@@ -269,12 +269,12 @@ public function partial(Dataset $dataset) : void
         }
 
         if ($this->fitPriors) {
-            $total = array_sum($this->weights) + (count($this->weights) * $this->alpha);
+            $total = array_sum($this->weights) + (count($this->weights) * $this->smoothing);
 
             $this->logPriors = [];
 
             foreach ($this->weights as $class => $weight) {
-                $this->logPriors[$class] = log(($weight + $this->alpha) / $total);
+                $this->logPriors[$class] = log(($weight + $this->smoothing) / $total);
             }
         }
     }

tests/Classifiers/NaiveBayesTest.php

@@ -135,7 +135,7 @@ public function compatibility() : void
     public function params() : void
     {
         $expected = [
-            'alpha' => 1.0,
+            'smoothing' => 1.0,
             'priors' => null,
         ];