PHP code example of jenutka / titanic_php

    
        
<?php
require_once('vendor/autoload.php');

/* Start to develop here. Best regards https://php-download.com/ */

    

jenutka / titanic_php example snippets

use Rubix\ML\Extractors\CSV;
use Rubix\ML\Extractors\ColumnPicker;

$extractor = new ColumnPicker(new CSV('train.csv', true), [
    'Pclass', 'Age', 'Fare', 'SibSp', 'Parch', 'Sex', 'Embarked', 'Survived',
]);

use Rubix\ML\Transformers\LambdaFunction;

$toPlaceholder = function (&$sample, $offset, $types) {
    foreach ($sample as $column => &$value) {
        if (empty($value) && $types[$column]->isContinuous()) {
            $value = NAN;
        }
        else if (empty($value) && $types[$column]->isCategorical()) {
            $value = '?';
        }
    }
};

$transformLabel = function ($label) {
    return $label == 0 ? 'Dead' : 'Survived';
};

use Rubix\ML\Transformers\MinMaxNormalizer;
use Rubix\ML\Transformers\OneHotEncoder;
use Rubix\ML\Transformers\MissingDataImputer;

$minMaxNormalizer = new MinMaxNormalizer();
$oneHotEncoder = new OneHotEncoder();
$imputer = new MissingDataImputer();

use Rubix\ML\Datasets\Labeled;

$dataset = Labeled::fromIterator($extractor)
    ->apply(new NumericStringConverter())
    ->transformLabels($transformLabel);

$dataset->apply(new LambdaFunction($toPlaceholder, $dataset->types()))
    ->apply($imputer)
    ->apply($minMaxNormalizer)
    ->apply($oneHotEncoder);

use Rubix\ML\Persisters\Filesystem;
use Rubix\ML\Serializers\RBX;

$serializer->serialize($imputer)->saveTo(new Filesystem('imputer.rbx'));
$serializer->serialize($minMaxNormalizer)->saveTo(new Filesystem('minmax.rbx'));
$serializer->serialize($oneHotEncoder)->saveTo(new Filesystem('onehot.rbx'));

use Rubix\ML\Classifiers\RandomForest;
use Rubix\ML\Classifiers\ClassificationTree;

$estimator = new RandomForest(new ClassificationTree(10), 500, 0.8, false);

$estimator->train($dataset);

use Rubix\ML\PersistentModel;

if (strtolower(readline('Save this model? (y|[n]): ')) === 'y') {
    $estimator = new PersistentModel($estimator, new Filesystem('model.rbx'));

    $estimator->save();

    $logger->info('Model saved as model.rbx');
}

use Rubix\ML\Extractors\ColumnPicker;

$extractor = new ColumnPicker(new CSV('test.csv', true), [
    'Pclass', 'Age', 'Fare', 'SibSp', 'Parch', 'Sex', 'Embarked',
]);

$persister_imputer = new Filesystem('imputer.rbx', true, new RBX());

$imputer = $persister_imputer->load()->deserializeWith(new RBX);

$persister_minMax = new Filesystem('minmax.rbx', true, new RBX());

$minMaxNormalizer = $persister_minMax->load()->deserializeWith(new RBX);

$persister_oneHot = new Filesystem('onehot.rbx', true, new RBX());

$oneHotEncoder = $persister_oneHot->load()->deserializeWith(new RBX);

use Rubix\ML\Datasets\Unlabeled;
use Rubix\ML\Transformers\LambdaFunction;

$dataset = Unlabeled::fromIterator($extractor)
    ->apply(new NumericStringConverter());

$dataset->apply(new LambdaFunction($toPlaceholder, $dataset->types()))
    ->apply($imputer)
    ->apply($minMaxNormalizer)
    ->apply($oneHotEncoder);

use Rubix\ML\PersistentModel;
use Rubix\ML\Persisters\Filesystem;
use Rubix\ML\Serializers\RBX;

$estimator = PersistentModel::load(new Filesystem('model.rbx'));

$predictions = $estimator->predict($dataset);

$predictions = $estimator->predict($dataset);

function bin_mapper($v)
{
    if ($v==="Survived") {
        return "1";
    } else {
        return "0";
    }
}

$predictions_mapped = array_map('bin_mapper', $predictions);

$extractor = new ColumnPicker(new CSV('test.csv', true), ['PassengerId']);

$ids = array_column(iterator_to_array($extractor), 'PassengerId');

array_unshift($ids, 'PassengerId');
array_unshift($predictions_mapped, 'Survived');

$extractor = new CSV('predictions.csv');

$extractor->export(array_transpose([$ids, $predictions_mapped]));

sh
$ composer create-project jenutka/titanic_php

sh
$ php train.php

sh
$ php predict.php