From 65cdfe64b2c307b223fd90f160e49d90e9cc4321 Mon Sep 17 00:00:00 2001
From: Arkadiusz Kondas <arkadiusz.kondas@gmail.com>
Date: Sun, 8 May 2016 19:33:39 +0200
Subject: [PATCH] implement Median and MostFrequent strategy for imputer

---
 src/Phpml/Math/Statistic/Mean.php             |  5 +-
 .../Imputer/Strategy/MedianStrategy.php       | 21 ++++
 .../Imputer/Strategy/MostFrequentStrategy.php | 21 ++++
 tests/Phpml/Math/Statistic/MeanTest.php       |  2 -
 tests/Phpml/Preprocessing/ImputerTest.php     | 98 ++++++++++++++++++-
 5 files changed, 140 insertions(+), 7 deletions(-)
 create mode 100644 src/Phpml/Preprocessing/Imputer/Strategy/MedianStrategy.php
 create mode 100644 src/Phpml/Preprocessing/Imputer/Strategy/MostFrequentStrategy.php

diff --git a/src/Phpml/Math/Statistic/Mean.php b/src/Phpml/Math/Statistic/Mean.php
index 9d87a97..3804848 100644
--- a/src/Phpml/Math/Statistic/Mean.php
+++ b/src/Phpml/Math/Statistic/Mean.php
@@ -29,8 +29,8 @@ class Mean
      *
      * @throws InvalidArgumentException
      */
-    public static function median(array $numbers) {
-
+    public static function median(array $numbers)
+    {
         self::checkArrayLength($numbers);
 
         $count = count($numbers);
@@ -72,5 +72,4 @@ class Mean
             throw InvalidArgumentException::arrayCantBeEmpty();
         }
     }
-
 }
diff --git a/src/Phpml/Preprocessing/Imputer/Strategy/MedianStrategy.php b/src/Phpml/Preprocessing/Imputer/Strategy/MedianStrategy.php
new file mode 100644
index 0000000..3746760
--- /dev/null
+++ b/src/Phpml/Preprocessing/Imputer/Strategy/MedianStrategy.php
@@ -0,0 +1,21 @@
+<?php
+
+declare (strict_types = 1);
+
+namespace Phpml\Preprocessing\Imputer\Strategy;
+
+use Phpml\Preprocessing\Imputer\Strategy;
+use Phpml\Math\Statistic\Mean;
+
+class MedianStrategy implements Strategy
+{
+    /**
+     * @param array $currentAxis
+     * 
+     * @return float
+     */
+    public function replaceValue(array $currentAxis)
+    {
+        return Mean::median($currentAxis);
+    }
+}
diff --git a/src/Phpml/Preprocessing/Imputer/Strategy/MostFrequentStrategy.php b/src/Phpml/Preprocessing/Imputer/Strategy/MostFrequentStrategy.php
new file mode 100644
index 0000000..690a97c
--- /dev/null
+++ b/src/Phpml/Preprocessing/Imputer/Strategy/MostFrequentStrategy.php
@@ -0,0 +1,21 @@
+<?php
+
+declare (strict_types = 1);
+
+namespace Phpml\Preprocessing\Imputer\Strategy;
+
+use Phpml\Preprocessing\Imputer\Strategy;
+use Phpml\Math\Statistic\Mean;
+
+class MostFrequentStrategy implements Strategy
+{
+    /**
+     * @param array $currentAxis
+     * 
+     * @return float|mixed
+     */
+    public function replaceValue(array $currentAxis)
+    {
+        return Mean::mode($currentAxis);
+    }
+}
diff --git a/tests/Phpml/Math/Statistic/MeanTest.php b/tests/Phpml/Math/Statistic/MeanTest.php
index 289f9a3..805e6a7 100644
--- a/tests/Phpml/Math/Statistic/MeanTest.php
+++ b/tests/Phpml/Math/Statistic/MeanTest.php
@@ -8,7 +8,6 @@ use Phpml\Math\Statistic\Mean;
 
 class MeanTest extends \PHPUnit_Framework_TestCase
 {
-
     /**
      * @expectedException \Phpml\Exception\InvalidArgumentException
      */
@@ -61,5 +60,4 @@ class MeanTest extends \PHPUnit_Framework_TestCase
 
         $this->assertEquals(6, Mean::mode($numbers));
     }
-
 }
diff --git a/tests/Phpml/Preprocessing/ImputerTest.php b/tests/Phpml/Preprocessing/ImputerTest.php
index 630f4b1..9da9765 100644
--- a/tests/Phpml/Preprocessing/ImputerTest.php
+++ b/tests/Phpml/Preprocessing/ImputerTest.php
@@ -6,10 +6,12 @@ namespace tests\Preprocessing;
 
 use Phpml\Preprocessing\Imputer;
 use Phpml\Preprocessing\Imputer\Strategy\MeanStrategy;
+use Phpml\Preprocessing\Imputer\Strategy\MedianStrategy;
+use Phpml\Preprocessing\Imputer\Strategy\MostFrequentStrategy;
 
 class ImputerTest extends \PHPUnit_Framework_TestCase
 {
-    public function testCompletingMissingValuesWithMeanStrategyOnColumnAxis()
+    public function testComplementsMissingValuesWithMeanStrategyOnColumnAxis()
     {
         $data = [
             [1, null, 3, 4],
@@ -31,7 +33,7 @@ class ImputerTest extends \PHPUnit_Framework_TestCase
         $this->assertEquals($imputeData, $data, '', $delta = 0.01);
     }
 
-    public function testCompletingMissingValuesWithMeanStrategyOnRowAxis()
+    public function testComplementsMissingValuesWithMeanStrategyOnRowAxis()
     {
         $data = [
             [1, null, 3, 4],
@@ -52,4 +54,96 @@ class ImputerTest extends \PHPUnit_Framework_TestCase
 
         $this->assertEquals($imputeData, $data, '', $delta = 0.01);
     }
+
+    public function testComplementsMissingValuesWithMediaStrategyOnColumnAxis()
+    {
+        $data = [
+            [1, null, 3, 4],
+            [4, 3, 2, 1],
+            [null, 6, 7, 8],
+            [8, 7, null, 5],
+        ];
+
+        $imputeData = [
+            [1, 6, 3, 4],
+            [4, 3, 2, 1],
+            [4, 6, 7, 8],
+            [8, 7, 3, 5],
+        ];
+
+        $imputer = new Imputer(null, new MedianStrategy(), Imputer::AXIS_COLUMN);
+        $imputer->preprocess($data);
+
+        $this->assertEquals($imputeData, $data, '', $delta = 0.01);
+    }
+
+    public function testComplementsMissingValuesWithMediaStrategyOnRowAxis()
+    {
+        $data = [
+            [1, null, 3, 4],
+            [4, 3, 2, 1],
+            [null, 6, 7, 8],
+            [8, 7, null, 5],
+        ];
+
+        $imputeData = [
+            [1, 3, 3, 4],
+            [4, 3, 2, 1],
+            [7, 6, 7, 8],
+            [8, 7, 7, 5],
+        ];
+
+        $imputer = new Imputer(null, new MedianStrategy(), Imputer::AXIS_ROW);
+        $imputer->preprocess($data);
+
+        $this->assertEquals($imputeData, $data, '', $delta = 0.01);
+    }
+
+    public function testComplementsMissingValuesWithMostFrequentStrategyOnColumnAxis()
+    {
+        $data = [
+            [1, null, 3, 4],
+            [4, 3, 2, 1],
+            [null, 6, 7, 8],
+            [8, 7, null, 5],
+            [8, 3, 2, 5],
+        ];
+
+        $imputeData = [
+            [1, 3, 3, 4],
+            [4, 3, 2, 1],
+            [8, 6, 7, 8],
+            [8, 7, 2, 5],
+            [8, 3, 2, 5],
+        ];
+
+        $imputer = new Imputer(null, new MostFrequentStrategy(), Imputer::AXIS_COLUMN);
+        $imputer->preprocess($data);
+
+        $this->assertEquals($imputeData, $data);
+    }
+
+    public function testComplementsMissingValuesWithMostFrequentStrategyOnRowAxis()
+    {
+        $data = [
+            [1, null, 3, 4, 3],
+            [4, 3, 2, 1, 7],
+            [null, 6, 7, 8, 6],
+            [8, 7, null, 5, 5],
+            [8, 3, 2, 5, 4],
+        ];
+
+        $imputeData = [
+            [1, 3, 3, 4, 3],
+            [4, 3, 2, 1, 7],
+            [6, 6, 7, 8, 6],
+            [8, 7, 5, 5, 5],
+            [8, 3, 2, 5, 4],
+        ];
+
+        $imputer = new Imputer(null, new MostFrequentStrategy(), Imputer::AXIS_ROW);
+        $imputer->preprocess($data);
+
+        $this->assertEquals($imputeData, $data);
+    }
 }