Phpml\Classification\Ensemble\RandomForest

Complexity

Total Complexity

10

Size/Duplication

Total Lines	70
Duplicated Lines	0 %

Coupling/Cohesion

Components	1
Dependencies	1

Importance

Changes

0

3 Methods

Rating	Name	Duplication	Size	Complexity
A	setFeatureSubsetRatio()	0	8	3
B	getRandomSubset()	0	23	4
A	predictSample()	0	17	3

<?php
declare(strict_types=1);

namespace Phpml\Classification\Ensemble;

use Phpml\Classification\Ensemble\Bagging;
use Phpml\Classification\DecisionTree;
use Phpml\Classification\NaiveBayes;

class RandomForest extends Bagging
{
    /**
     * @var array
     */
    protected $classifierColumns;

    /**
     * @var float
     */
    protected $featureSubsetRatio = 0.7;

    public function setFeatureSubsetRatio(float $ratio)
    {
        if ($ratio < 0.1 || $ratio > 0.9) {
            throw new Exception("Feature subset ratio should be between 0.1 and 0.9");
        }
        $this->featureSubsetRatio = $ratio;
        return $this;
    }

    /**
     * @param int $index
     */
    protected function getRandomSubset($index)
    {
        list($subset, $targets) = parent::getRandomSubset($index);

        $features = [];
        $featureCount = (int)ceil($this->featureSubsetRatio * $this->featureCount);
        while (count($features) < $featureCount) {
            $rand = rand(0, $this->featureCount - 1);
            if (! in_array($rand, $features)) {
                $features[] = $rand;
            }
        }
        sort($features);
        $this->classifierColumns[$index] = $features;

        $columns = [];
        foreach ($features as $colIndex) {
            $columns[] = array_column($subset, $colIndex);
        }
        $subset= array_map(null, ...$columns);

        return [$subset, $targets];
    }

    /**
     * @param array $sample
     * @return mixed
     */
    protected function predictSample(array $sample)
    {
        $predictions = [];
        for ($i=0; $i<count($this->classifiers); $i++) {

It seems like you are calling the size function count() as part of the test condition. You might want to compute the size beforehand, and not on each iteration.

            $samplePiece = [];
            foreach ($this->classifierColumns[$i] as $colIndex) {
                $samplePiece[] = $sample[$colIndex];
            }
            /* @var $classifier Classifier */
            $predictions[] = $this->classifiers[$i]->predict($samplePiece);
        }

        $counts = array_count_values($predictions);
        arsort($counts);
        reset($counts);
        return key($counts);
    }
}