mirror of
https://github.com/Llewellynvdm/php-ml.git
synced 2024-11-29 08:13:59 +00:00
e255369636
* Update docs for Imputer class * Throw exception when trying to transform imputer without train data * Update changelog
68 lines
1.7 KiB
Markdown
68 lines
1.7 KiB
Markdown
# Imputation missing values
|
|
|
|
For various reasons, many real world datasets contain missing values, often encoded as blanks, NaNs or other placeholders.
|
|
To solve this problem you can use the `Imputer` class.
|
|
|
|
## Constructor Parameters
|
|
|
|
* $missingValue (mixed) - this value will be replaced (default null)
|
|
* $strategy (Strategy) - imputation strategy (read to use: MeanStrategy, MedianStrategy, MostFrequentStrategy)
|
|
* $axis (int) - axis for strategy, Imputer::AXIS_COLUMN or Imputer::AXIS_ROW
|
|
* $samples (array) - array of samples to train
|
|
|
|
```
|
|
$imputer = new Imputer(null, new MeanStrategy(), Imputer::AXIS_COLUMN);
|
|
$imputer = new Imputer(null, new MedianStrategy(), Imputer::AXIS_ROW);
|
|
```
|
|
|
|
## Strategy
|
|
|
|
* MeanStrategy - replace missing values using the mean along the axis
|
|
* MedianStrategy - replace missing values using the median along the axis
|
|
* MostFrequentStrategy - replace missing using the most frequent value along the axis
|
|
|
|
## Example of use
|
|
|
|
```
|
|
use Phpml\Preprocessing\Imputer;
|
|
use Phpml\Preprocessing\Imputer\Strategy\MeanStrategy;
|
|
|
|
$data = [
|
|
[1, null, 3, 4],
|
|
[4, 3, 2, 1],
|
|
[null, 6, 7, 8],
|
|
[8, 7, null, 5],
|
|
];
|
|
|
|
$imputer = new Imputer(null, new MeanStrategy(), Imputer::AXIS_COLUMN);
|
|
$imputer->fit($data);
|
|
$imputer->transform($data);
|
|
|
|
/*
|
|
$data = [
|
|
[1, 5.33, 3, 4],
|
|
[4, 3, 2, 1],
|
|
[4.33, 6, 7, 8],
|
|
[8, 7, 4, 5],
|
|
];
|
|
*/
|
|
|
|
```
|
|
|
|
You can also use `$samples` constructer parameter instead of `fit` method:
|
|
|
|
```
|
|
use Phpml\Preprocessing\Imputer;
|
|
use Phpml\Preprocessing\Imputer\Strategy\MeanStrategy;
|
|
|
|
$data = [
|
|
[1, null, 3, 4],
|
|
[4, 3, 2, 1],
|
|
[null, 6, 7, 8],
|
|
[8, 7, null, 5],
|
|
];
|
|
|
|
$imputer = new Imputer(null, new MeanStrategy(), Imputer::AXIS_COLUMN, $data);
|
|
$imputer->transform($data);
|
|
```
|