[Summary view]

Differences Between: [Versions 311 and 400] [Versions 311 and 401] [Versions 311 and 402] [Versions 311 and 403]

   1  <?php
   2  
   3  declare(strict_types=1);
   4  
   5  namespace Phpml\Preprocessing;
   6  
   7  use Phpml\Exception\InvalidOperationException;
   8  use Phpml\Preprocessing\Imputer\Strategy;
   9  
  10  class Imputer implements Preprocessor
  11  {
  12      public const AXIS_COLUMN = 0;
  13  
  14      public const AXIS_ROW = 1;
  15  
  16      /**
  17       * @var mixed
  18       */
  19      private $missingValue;
  20  
  21      /**
  22       * @var Strategy
  23       */
  24      private $strategy;
  25  
  26      /**
  27       * @var int
  28       */
  29      private $axis;
  30  
  31      /**
  32       * @var mixed[]
  33       */
  34      private $samples = [];
  35  
  36      /**
  37       * @param mixed $missingValue
  38       */
  39      public function __construct($missingValue, Strategy $strategy, int $axis = self::AXIS_COLUMN, array $samples = [])
  40      {
  41          $this->missingValue = $missingValue;
  42          $this->strategy = $strategy;
  43          $this->axis = $axis;
  44          $this->samples = $samples;
  45      }
  46  
  47      public function fit(array $samples, ?array $targets = null): void
  48      {
  49          $this->samples = $samples;
  50      }
  51  
  52      public function transform(array &$samples): void
  53      {
  54          if ($this->samples === []) {
  55              throw new InvalidOperationException('Missing training samples for Imputer.');
  56          }
  57  
  58          foreach ($samples as &$sample) {
  59              $this->preprocessSample($sample);
  60          }
  61      }
  62  
  63      private function preprocessSample(array &$sample): void
  64      {
  65          foreach ($sample as $column => &$value) {
  66              if ($value === $this->missingValue) {
  67                  $value = $this->strategy->replaceValue($this->getAxis($column, $sample));
  68              }
  69          }
  70      }
  71  
  72      private function getAxis(int $column, array $currentSample): array
  73      {
  74          if ($this->axis === self::AXIS_ROW) {
  75              return array_diff($currentSample, [$this->missingValue]);
  76          }
  77  
  78          $axis = [];
  79          foreach ($this->samples as $sample) {
  80              if ($sample[$column] !== $this->missingValue) {
  81                  $axis[] = $sample[$column];
  82              }
  83          }
  84  
  85          return $axis;
  86      }
  87  }