### Example: Setting and Using Screen Logger

Source: https://github.com/rubixml/ml/blob/master/docs/verbose.md

This example demonstrates how to instantiate an Adaline regressor, set a Screen logger for real-time output, and then train the estimator. The output shows the logger capturing initialization, training progress, and completion messages.

```php
use Rubix\ML\Regressors\Adaline;
use Rubix\ML\Loggers\Screen;

$estimator = new Adaline();

$estimator->setLogger(new Screen('example'));

$estimator->train($dataset);
```

```text
[2020-08-05 04:26:11] INFO: Learner init Adaline {batch_size: 128, optimizer: Adam {rate: 0.01, momentum_decay: 0.1, norm_decay: 0.001}, alpha: 0.0001, epochs: 100, min_change: 0.001, window: 5, cost_fn: Huber Loss {alpha: 1}}
[2020-08-05 04:26:11] INFO: Training started
[2020-08-05 04:26:11] example.INFO: Epoch 1 - Huber Loss {alpha: 1}: 0.36839299586132
[2020-08-05 04:26:11] example.INFO: Epoch 2 - Huber Loss {alpha: 1}: 0.0018235958273629
[2020-08-05 04:26:11] example.INFO: Epoch 3 - Huber Loss {alpha: 1}: 0.0017358090553563
[2020-08-05 04:26:11] example.INFO: Training complete
```

--------------------------------

### Install Python Dependencies for Docs

Source: https://github.com/rubixml/ml/blob/master/CONTRIBUTING.md

Install necessary Python packages for compiling and managing documentation with Mkdocs and Mike.

```sh
$ pip install mike mkdocs mkdocs-material mkdocs-git-revision-date-localized-plugin
```

--------------------------------

### Training Progress Output Example

Source: https://github.com/rubixml/ml/blob/master/docs/training.md

Example output from the Screen logger during the training of a Logistic Regression model, showing epoch-wise loss progression and final training completion.

```text
[2020-09-04 08:39:04] INFO: Logistic Regression (batch_size: 128, optimizer: Adam (rate: 0.01, momentum_decay: 0.1, norm_decay: 0.001), alpha: 0.0001, epochs: 1000, min_change: 0.0001, window: 5, cost_fn: Cross Entropy) initialized
[2020-09-04 08:39:04] INFO: Epoch 1 - Cross Entropy: 0.16895133388673
[2020-09-04 08:39:04] INFO: Epoch 2 - Cross Entropy: 0.16559247705179
[2020-09-04 08:39:04] INFO: Epoch 3 - Cross Entropy: 0.16294448401323
[2020-09-04 08:39:04] INFO: Epoch 4 - Cross Entropy: 0.16040135038265
[2020-09-04 08:39:04] INFO: Epoch 5 - Cross Entropy: 0.15786801071483
[2020-09-04 08:39:04] INFO: Epoch 6 - Cross Entropy: 0.1553151426337
[2020-09-04 08:39:04] INFO: Epoch 7 - Cross Entropy: 0.15273253982757
[2020-09-04 08:39:04] INFO: Epoch 8 - Cross Entropy: 0.15011771931339
[2020-09-04 08:39:04] INFO: Epoch 9 - Cross Entropy: 0.14747194148672
[2020-09-04 08:39:04] INFO: Epoch 10 - Cross Entropy: 0.14479847759871
...
[2020-09-04 08:39:04] INFO: Epoch 77 - Cross Entropy: 0.0082096137827592
[2020-09-04 08:39:04] INFO: Epoch 78 - Cross Entropy: 0.0081004235278088
[2020-09-04 08:39:04] INFO: Epoch 79 - Cross Entropy: 0.0079956096838174
[2020-09-04 08:39:04] INFO: Epoch 80 - Cross Entropy: 0.0078948616067878
[2020-09-04 08:39:04] INFO: Epoch 81 - Cross Entropy: 0.0077978960869396
[2020-09-04 08:39:04] INFO: Training complete
```

--------------------------------

### Representing Samples and Labels in PHP

Source: https://github.com/rubixml/ml/blob/master/docs/basic-introduction.md

This example shows how to structure your data for machine learning tasks in PHP using a 2-d array for samples and a 1-d array for labels.

```php
$samples = [
	[3, 4, 50.5],
	[1, 5, 24.7],
	[4, 4, 62.0],
	[3, 2, 31.1],
];

$labels = ['married', 'divorced', 'married', 'divorced'];
```

--------------------------------

### Initialize MultibyteTextNormalizer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/multibyte-text-normalizer.md

Instantiate the MultibyteTextNormalizer to convert text to lowercase. Ensure the mbstring extension is installed for best performance.

```php
use Rubix\ML\Transformers\MultibyteTextNormalizer;

$transformer = new MultibyteTextNormalizer(false);
```

--------------------------------

### Instantiate a K-Nearest Neighbors Classifier

Source: https://github.com/rubixml/ml/blob/master/docs/choosing-an-estimator.md

Instantiate a KNearestNeighbors classifier with custom hyper-parameters. This example shows how to set the number of neighbors, enable/disable probability estimation, and specify a distance kernel.

```php
use Rubix\ML\Classifiers\KNearestNeighbors;
use Rubix\ML\Kernels\Distance\Minkowski;

$estimator = new KNearestNeighbors(10, false, new Minkowski(2.5));
```

--------------------------------

### Initialize PlusPlus Seeder with Minkowski Kernel

Source: https://github.com/rubixml/ml/blob/master/docs/clusterers/seeders/plus-plus.md

Instantiate the PlusPlus seeder with a custom distance kernel. This example uses the Minkowski distance with a power of 5.0.

```php
use Rubix\ML\Clusterers\Seeders\PlusPlus;
use Rubix\ML\Kernels\Distance\Minkowski;

$seeder = new PlusPlus(new Minkowski(5.0));
```

--------------------------------

### Set Parallel Backend

Source: https://github.com/rubixml/ml/blob/master/docs/parallel.md

Configure a parallelizable object to use a specific backend for parallel processing. This example demonstrates setting the Amp backend with 16 processes.

```php
public setBackend(Backend $backend) : void
```

```php
use Rubix\ML\Classifiers\RandomForest;
use Rubix\ML\Backends\Amp;

$estimator = new RandomForest();

$estimator->setBackend(new Amp(16));
```

--------------------------------

### Example Predictions Output

Source: https://github.com/rubixml/ml/blob/master/docs/basic-introduction.md

The output of the predict() method are the predicted class labels for the provided unknown samples.

```php
Array
(
    [0] => married
    [1] => divorced
    [2] => divorced
    [3] => married
)
```

--------------------------------

### Instantiate RadiusNeighborsRegressor

Source: https://github.com/rubixml/ml/blob/master/docs/regressors/radius-neighbors-regressor.md

Instantiates the RadiusNeighborsRegressor with custom parameters. This example shows how to set the radius, disable weighted averaging, and specify a BallTree with a maximum leaf size of 30 and a Diagonal distance kernel.

```php
use Rubix\ML\Regressors\RadiusNeighborsRegressor;
use Rubix\ML\Graph\Trees\BallTree;
use Rubix\ML\Kernels\Distance\Diagonal;

$estimator = new RadiusNeighborsRegressor(0.5, false, new BallTree(30, new Diagonal()));
```

--------------------------------

### Initialize K Nearest Neighbors Classifier

Source: https://github.com/rubixml/ml/blob/master/docs/classifiers/k-nearest-neighbors.md

Instantiate the KNearestNeighbors classifier with custom parameters for k, weighting, and the distance kernel. This example uses the Manhattan distance kernel.

```php
use Rubix\ML\Classifiers\KNearestNeighbors;
use Rubix\ML\Kernels\Distance\Manhattan;

$estimator = new KNearestNeighbors(3, false, new Manhattan());
```

--------------------------------

### PHP Constant Initializer Example

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/initializers/constant.md

Instantiates the Constant initializer with a specific float value. This is used to set the initial weights or biases of a neural network layer.

```php
use Rubix\ML\NeuralNet\Initializers\Constant;

$initializer = new Constant(1.0);
```

--------------------------------

### Run PHP Script from Command Line

Source: https://github.com/rubixml/ml/blob/master/docs/faq.md

Execute a PHP script using the command line interface. Ensure the PHP interpreter is installed and in your system's PATH.

```sh
$ php example.php
```

--------------------------------

### Instantiate KDTree with Euclidean Kernel

Source: https://github.com/rubixml/ml/blob/master/docs/graph/trees/k-d-tree.md

Creates a new KDTree instance with a maximum leaf size of 30 and the Euclidean distance kernel. This is a common starting point for using the KDTree for spatial data.

```php
use Rubix\ML\Graph\Trees\KDTree;
use Rubix\ML\Kernels\Distance\Euclidean;

$tree = new KDTree(30, new Euclidean());
```

--------------------------------

### Instantiate ReLU Activation Function

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/activation-functions/relu.md

Instantiates the ReLU activation function. Note that the constructor signature in the example might differ from the actual implementation if default parameters are not explicitly shown.

```php
use Rubix\ML\NeuralNet\ActivationFunctions\ReLU;

$activationFunction = new ReLU(0.1);
```

--------------------------------

### Install Rubix ML with Composer

Source: https://github.com/rubixml/ml/blob/master/README.md

Install Rubix ML into your project using Composer. Ensure you have PHP 7.4 or above installed.

```sh
composer require rubix/ml
```

--------------------------------

### Transformer Pipeline for Sequential Transformations

Source: https://github.com/rubixml/ml/blob/master/docs/preprocessing.md

Automate a series of transformations using the Pipeline meta-estimator. This example sets up a pipeline with imputation, encoding, and standardization before passing data to a KMeans clusterer.

```php
use Rubix\ML\Pipeline;
use Rubix\ML\Transformers\HotDeckImputer;
use Rubix\ML\Transformers\OneHotEncoder;
use Rubix\ML\Transformers\ZScaleStandardizer;
use Rubix\ML\Clusterers\KMeans;

$estimator = new Pipeline([
    new HotDeckImputer(5),
    new OneHotEncoder(),
    new ZScaleStandardizer(),
], new KMeans(10));
```

--------------------------------

### Deploy and Serve Documentation

Source: https://github.com/rubixml/ml/blob/master/CONTRIBUTING.md

Deploy documentation for a specific version and serve the documentation site locally for development.

```sh
$ mike deploy 'VERSION'
```

```sh
$ mike serve
```

--------------------------------

### Get Class Revision Hash

Source: https://github.com/rubixml/ml/blob/master/docs/persistable.md

Call the revision method on a Persistable object to get its current class revision hash.

```php
echo $persistable->revision();
```

--------------------------------

### Initialize Grid Search with KNearestNeighbors

Source: https://github.com/rubixml/ml/blob/master/docs/grid-search.md

Instantiate Grid Search with a base learner class, parameter combinations, a validation metric, and a validator.

```php
use Rubix\ML\GridSearch;
use Rubix\ML\Classifiers\KNearestNeighbors;
use Rubix\ML\Kernels\Distance\Euclidean;
use Rubix\ML\Kernels\Distance\Manhattan;
use Rubix\ML\CrossValidation\Metrics\FBeta;
use Rubix\ML\CrossValidation\KFold;

$params = [
	[1, 3, 5, 10], [true, false], [new Euclidean(), new Manhattan()],
];

$estimator = new GridSearch(KNearestNeighbors::class, $params, new FBeta(), new KFold(5));
```

--------------------------------

### Example Accuracy Score

Source: https://github.com/rubixml/ml/blob/master/docs/basic-introduction.md

This is an example of the accuracy score returned by the test method, indicating the model's performance on unseen data.

```text
0.88
```

--------------------------------

### Initialize Word Tokenizer

Source: https://github.com/rubixml/ml/blob/master/docs/tokenizers/word.md

Instantiate the Word tokenizer. This tokenizer does not accept any parameters.

```php
use Rubix\ML\Tokenizers\Word;

$tokenizer = new Word();
```

--------------------------------

### Advanced Preprocessing with Column Pickers and Pipelines

Source: https://github.com/rubixml/ml/blob/master/docs/preprocessing.md

Extract specific features using Column Pickers, apply distinct transformations to different feature sets, and then join them. This example preprocesses text reviews and categorical features separately before joining and applying a standardizer.

```php
use Rubix\ML\Dataset\Labeled;
use Rubix\ML\Extractors\ColumnPicker;
use Rubix\ML\Extractors\NDJSON;
use Rubix\ML\Dataset\Unlabeled;
use Rubix\ML\Transformers\TextNormalizer;
use Rubix\ML\Transformers\WordCountVectorizer;
use Rubix\ML\Transformers\TfIdfTransformer;
use Rubix\ML\Transformers\OneHotEncoder;
use Rubix\ML\Transformers\ZScaleStandardizer;

$extractor1 = new ColumnPicker(new NDJSON('example.ndjson'), [
    'review', 'sentiment',
]);

$extractor2 = new ColumnPicker(new NDJSON('example.ndjson'), [
    'category', 'clicks', 'rating',
]);

$dataset1 = Labeled::fromIterator($extractor1)
    ->apply(new TextNormalizer())
    ->apply(new WordCountVectorizer(5000))
    ->apply(new TfIdfTransformer());

$dataset2 = Unlabeled::fromIterator($extractor2)
    ->apply(new OneHotEncoder());

$dataset = $dataset1->join($dataset2)
    ->apply(new ZScaleStandardizer());
```

--------------------------------

### Instantiate Screen Logger

Source: https://github.com/rubixml/ml/blob/master/docs/loggers/screen.md

Create a new instance of the Screen logger. You can specify a channel name and a timestamp format.

```php
use Rubix\ML\Loggers\Screen;

$logger = new Screen('mlp', 'Y-m-d H:i:s');
```

--------------------------------

### Contingency Table JSON Output Example

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/reports/contingency-table.md

Example of the JSON output format for a Contingency Table report, illustrating the frequency distribution of class labels (e.g., 'lamb', 'wolf') across different clusters.

```json
[
    {
        "lamb": 11,
        "wolf": 2
    },
    {
        "lamb": 1,
        "wolf": 5
    }
]
```

--------------------------------

### Instantiate Softmax Activation Function

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/activation-functions/softmax.md

Learn how to create an instance of the Softmax activation function. No parameters are required for initialization.

```php
use Rubix\ML\NeuralNet\ActivationFunctions\Softmax;

$activationFunction = new Softmax();
```

--------------------------------

### Instantiate Serial Backend

Source: https://github.com/rubixml/ml/blob/master/docs/backends/serial.md

Create an instance of the Serial backend. This backend is suitable for sequential task execution and is often the default choice for small datasets due to its zero overhead.

```php
use Rubix\ML\Backends\Serial;

$backend = new Serial();
```

--------------------------------

### Get Labels

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Retrieves all the labels from the dataset as an array.

```APIDOC
## labels()

### Description
Return the labels of the dataset in an array.

### Method
`public labels() : array`

### Parameters
None

### Request Example
```php
$dataset->labels();
```

### Response
#### Success Response (200)
An array containing all labels.

#### Response Example
```php
[
    'not monster',
    'monster',
    'not monster'
]
```
```

--------------------------------

### Get Label Type

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Returns the data type of the labels in the dataset.

```APIDOC
## labelType()

### Description
Return the data type of the label.

### Method
`public labelType() : Rubix\ML\DataType`

### Parameters
None

### Request Example
```php
echo $dataset->labelType();
```

### Response
#### Success Response (200)
The data type of the label.

#### Response Example
```sh
continuous
```
```

--------------------------------

### Get Possible Outcomes

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Returns an array of all unique labels present in the dataset.

```APIDOC
## possibleOutcomes()

### Description
Return all of the possible outcomes i.e. the unique labels in an array.

### Method
`public possibleOutcomes() : array`

### Parameters
None

### Request Example
```php
$dataset->possibleOutcomes();
```

### Response
#### Success Response (200)
An array of unique labels.

#### Response Example
```php
[
    'female',
    'male'
]
```
```

--------------------------------

### Run Benchmarking Suite

Source: https://github.com/rubixml/ml/blob/master/CONTRIBUTING.md

Execute the benchmarking suite using Composer to assess performance.

```sh
$ composer benchmark
```

--------------------------------

### Get Label by Offset

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Retrieves a single label at a specified row offset.

```APIDOC
## label()

### Description
Return a single label at the given row offset.

### Method
`public label(int $offset) : mixed`

### Parameters
#### Path Parameters
* **offset** (int) - Required - The row offset of the label to retrieve.

### Request Example
```php
$dataset->label(1);
```

### Response
#### Success Response (200)
The label at the specified offset.

#### Response Example
```php
'monster'
```
```

--------------------------------

### Initialize HotDeckImputer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/hot-deck-imputer.md

Instantiates the HotDeckImputer with specified parameters including k, weighted, categoricalPlaceholder, and a BallTree with a Gower distance kernel.

```php
use Rubix\ML\Transformers\HotDeckImputer;
use Rubix\ML\Graph\Trees\BallTree;
use Rubix\ML\Kernels\Distance\Gower;

$transformer = new HotDeckImputer(20, false, '?', new BallTree(50, new Gower(1.0)));
```

--------------------------------

### Get Descriptive Statistics

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Retrieves descriptive statistics for continuous and categorical features in the dataset.

```php
echo $dataset->describe();
```

```json
[
    {
        "offset": 0,
        "type": "categorical",
        "num categories": 2,
        "probabilities": {
            "friendly": 0.6666666666666666,
            "loner": 0.3333333333333333
        }
    },
    {
        "offset": 1,
        "type": "continuous",
        "mean": 0.3333333333333333,
        "standard deviation": 3.129252661934191,
        "skewness": -0.4481030843690633,
        "kurtosis": -1.1330702741786107,
        "min": -5,
        "25%": -1.375,
        "median": 0.8,
        "75%": 2.825,
        "max": 4
    }
]
```

--------------------------------

### splice

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Removes a size *n* chunk of the dataset starting at *offset* and returns it in a new dataset.

```APIDOC
## splice(int $offset, int $n)

### Description
Remove a size *n* chunk of the dataset starting at *offset* and return it in a new dataset.

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **offset** (int) - Required - The starting offset of the chunk to remove.
- **n** (int) - Required - The size of the chunk to remove.

### Response
#### Success Response (200)
- **return value** (self) - A new dataset object containing the removed chunk.
```

--------------------------------

### Instantiate Pipeline with Transformers and Estimator

Source: https://github.com/rubixml/ml/blob/master/docs/pipeline.md

Instantiate a Pipeline with a list of transformers, a base estimator, and the elastic mode setting. This is useful for creating a preprocessing and modeling workflow.

```php
use Rubix\ML\Pipeline;
use Rubix\ML\Transformers\MissingDataImputer;
use Rubix\ML\Transformers\OneHotEncoder;
use Rubix\ML\Transformers\PrincipalComponentAnalysis;
use Rubix\ML\Classifiers\SoftmaxClassifier;

$estimator = new Pipeline([
	new MissingDataImputer(),
	new OneHotEncoder(), 
	new PrincipalComponentAnalysis(20),
], new SoftmaxClassifier(128), true);
```

--------------------------------

### Instantiate Softsign Activation Function

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/activation-functions/softsign.md

This snippet shows how to create an instance of the Softsign activation function using PHP. No parameters are required for initialization.

```php
use Rubix\ML\NeuralNet\ActivationFunctions\Softsign;

$activationFunction = new Softsign();
```

--------------------------------

### Get Number of Imported Records

Source: https://github.com/rubixml/ml/blob/master/docs/extracting-data.md

Check the total number of records imported into the dataset object after extraction.

```php
echo $dataset->numSamples();
```

--------------------------------

### Instantiate Accuracy Metric

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/metrics/accuracy.md

Learn how to create an instance of the Accuracy metric. This is the first step before using it to evaluate models.

```php
use Rubix\ML\CrossValidation\Metrics\Accuracy;

$metric = new Accuracy();
```

--------------------------------

### Initialize Uniform Distribution

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/initializers/uniform.md

Instantiate the Uniform initializer with a beta value to set the bounds for the random weight distribution.

```php
use Rubix\ML\NeuralNet\Initializers\Uniform;

$initializer = new Uniform(1e-3);
```

--------------------------------

### Instantiate SiLU Activation Function

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/activation-functions/silu.md

This snippet shows how to create an instance of the SiLU activation function. No parameters are required.

```php
use Rubix\ML\NeuralNet\ActivationFunctions\SiLU;

$activationFunction = new SiLU();
```

--------------------------------

### Get Feature Importances

Source: https://github.com/rubixml/ml/blob/master/docs/ranks-features.md

Retrieves the importance scores for each feature column from the training set. This method is called after the model has been trained.

```php
public featureImportances() : array
```

```php
$estimator->train($dataset);

$importances = $estimator->featureImportances();

print_r($importances);
```

```php
Array
(
    [0] => 0.04757
    [1] => 0.37948
    [2] => 0.53170
    [3] => 0.04123
)
```

--------------------------------

### Instantiate TruncatedSVD Transformer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/truncated-svd.md

Instantiate the TruncatedSVD transformer with a specified number of dimensions. The Tensor extension must be installed to use this transformer.

```php
use Rubix\ML\Transformers\TruncatedSVD;

$transformer = new TruncatedSVD(100);
```

--------------------------------

### Initialize Preset Seeder

Source: https://github.com/rubixml/ml/blob/master/docs/clusterers/seeders/preset.md

Instantiate the Preset seeder with an array of predefined centroids. Each centroid can be an array of values representing its features.

```php
use Rubix\ML\Clusterers\Seeders\Preset;

$seeder = new Preset([
    ['foo', 14, 0.72],
    ['bar', 16, 0.92],
]);
```

--------------------------------

### Score a Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/scoring.md

Returns the anomaly scores assigned to the samples in a dataset. Use this method to get anomaly scores for each data point.

```php
public score(Dataset $dataset) : array
```

```php
$scores = $estimator->score($dataset);

print_r($scores);
```

```php
Array
(
    [0] => 0.35033
    [1] => 0.40992
    [2] => 1.68153
)
```

--------------------------------

### Get Label Type

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Determine and display the data type of the labels in the dataset. This helps in understanding how labels are treated by the ML algorithms.

```php
echo $dataset->labelType();
```

--------------------------------

### Get Unique Labels

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Retrieve an array containing all unique class labels present in the dataset. This is useful for understanding the distribution of outcomes.

```php
print_r($dataset->possibleOutcomes());
```

--------------------------------

### Instantiate Amp Backend

Source: https://github.com/rubixml/ml/blob/master/docs/backends/amp.md

Instantiates the Amp backend with a specified number of workers. The number of workers can be auto-detected if not provided.

```php
use Rubix\ML\Backends\Amp;

$backend = new Amp(16);
```

--------------------------------

### Apply a Transformer to a Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/api.md

Pass a transformer object to the apply() method on a Dataset object to transform its features. This example uses MinMaxNormalizer.

```php
use Rubix\ML\Transformers\MinMaxNormalizer;

$dataset->apply(new MinMaxNormalizer());
```

--------------------------------

### Instantiate He Initializer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/initializers/he.md

This snippet shows how to instantiate the He initializer. It does not require any parameters.

```php
use Rubix\ML\NeuralNet\Initializers\He;

$initializer = new He();
```

--------------------------------

### Initialize PReLU Layer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/hidden-layers/prelu.md

Instantiate a PReLU layer with a custom initializer for the leakage parameter. This example uses the Normal initializer with a mean of 0.5.

```php
use Rubix\ML\NeuralNet\Layers\PReLU;
use Rubix\ML\NeuralNet\Initializers\Normal;

$layer = new PReLU(new Normal(0.5));
```

--------------------------------

### Batch vs. Online Learning

Source: https://github.com/rubixml/ml/blob/master/docs/training.md

Demonstrates batch learning using `train()` followed by online learning updates using `partial()` with subsequent datasets. Online learning is useful for large datasets or streaming data.

```php
$estimator->train($dataset1);

$estimator->partial($dataset2);

$estimator->partial($dataset3);
```

--------------------------------

### Get First N Rows

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Extract the first 'n' rows from the dataset, returning them as a new dataset object. This is useful for previewing the beginning of the data.

```php
$subset = $dataset->head(10);
```

--------------------------------

### Instantiate Vantage Tree

Source: https://github.com/rubixml/ml/blob/master/docs/graph/trees/vantage-tree.md

Instantiates a new Vantage Tree with a specified maximum leaf size and a Euclidean distance kernel. Ensure the Euclidean kernel is imported.

```php
use Rubix\ML\Graph\Trees\VantageTree;
use Rubix\ML\Kernels\Distance\Euclidean;

$tree = new VantageTree(30, new Euclidean());
```

--------------------------------

### Initialize Sparse Cosine Kernel

Source: https://github.com/rubixml/ml/blob/master/docs/kernels/distance/sparse-cosine.md

Instantiate the SparseCosine kernel. This kernel is optimized for sparse vectors and requires no parameters.

```php
use Rubix\ML\Kernels\Distance\SparseCosine;

$kernel = new SparseCosine();
```

--------------------------------

### Get Feature Type by Offset

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Retrieve the data type for a specific feature column using its offset. This helps in understanding the nature of the data in that column.

```php
echo $dataset->featureType(15);
```

--------------------------------

### Get Dataset Shape

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Retrieve the dimensions (number of rows and columns) of the samples matrix within the dataset. This is useful for understanding the dataset's size.

```php
[$m, $n] = $dataset->shape();

echo "$m x $n";
```

--------------------------------

### Initialize KNNImputer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/knn-imputer.md

Instantiates the KNNImputer with specified parameters for k, weighting, categorical placeholder, and a BallTree with a SafeEuclidean distance kernel.

```php
use Rubix\ML\Transformers\KNNImputer;
use Rubix\ML\Graph\Trees\BallTee;
use Rubix\ML\Kernels\Distance\SafeEuclidean;

$transformer = new KNNImputer(10, false, '?', new BallTree(30, new SafeEuclidean()));
```

--------------------------------

### Multiclass Breakdown Report JSON Output

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/reports/multiclass-breakdown.md

An example of the JSON output from the MulticlassBreakdown report, showing overall metrics and per-class metrics for a classification task.

```json
{
    "overall": {
        "accuracy": 0.6,
        "accuracy balanced": 0.5833333333333333,
        "f1 score": 0.5833333333333333,
        "precision": 0.5833333333333333,
        "recall": 0.5833333333333333,
        "specificity": 0.5833333333333333,
        "negative predictive value": 0.5833333333333333,
        "false discovery rate": 0.4166666666666667,
        "miss rate": 0.4166666666666667,
        "fall out": 0.4166666666666667,
        "false omission rate": 0.4166666666666667,
        "mcc": 0.16666666666666666,
        "informedness": 0.16666666666666652,
        "markedness": 0.16666666666666652,
        "true positives": 3,
        "true negatives": 3,
        "false positives": 2,
        "false negatives": 2,
        "cardinality": 5
    },
    "classes": {
        "wolf": {
            "accuracy": 0.6,
            "accuracy balanced": 0.5833333333333333,
            "f1 score": 0.6666666666666666,
            "precision": 0.6666666666666666,
            "recall": 0.6666666666666666,
            "specificity": 0.5,
            "negative predictive value": 0.5,
            "false discovery rate": 0.33333333333333337,
            "miss rate": 0.33333333333333337,
            "fall out": 0.5,
            "false omission rate": 0.5,
            "informedness": 0.16666666666666652,
            "markedness": 0.16666666666666652,
            "mcc": 0.16666666666666666,
            "true positives": 2,
            "true negatives": 1,
            "false positives": 1,
            "false negatives": 1,
            "cardinality": 3,
            "proportion": 0.6
        },
        "lamb": {
            "accuracy": 0.6,
            "accuracy balanced": 0.5833333333333333,
            "f1 score": 0.5,
            "precision": 0.5,
            "recall": 0.5,
            "specificity": 0.6666666666666666,
            "negative predictive value": 0.6666666666666666,
            "false discovery rate": 0.5,
            "miss rate": 0.5,
            "fall out": 0.33333333333333337,
            "false omission rate": 0.33333333333333337,
            "informedness": 0.16666666666666652,
            "markedness": 0.16666666666666652,
            "mcc": 0.16666666666666666,
            "true positives": 1,
            "true negatives": 2,
            "false positives": 1,
            "false negatives": 1,
            "cardinality": 2,
            "proportion": 0.4
        }
    }
}
```

--------------------------------

### Instantiate Linear SVM Kernel

Source: https://github.com/rubixml/ml/blob/master/docs/kernels/svm/linear.md

Learn how to create an instance of the Linear kernel for SVM models. This kernel requires no parameters.

```php
use Rubix\ML\Kernels\SVM\Linear;

$kernel = new Linear();
```

--------------------------------

### Instantiate Step Decay Optimizer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/optimizers/step-decay.md

Initializes the Step Decay optimizer with a learning rate of 0.1, a step interval of 50, and a decay factor of 1e-3.

```php
use Rubix\ML\NeuralNet\Optimizers\StepDecay;

$optimizer = new StepDecay(0.1, 50, 1e-3);
```

--------------------------------

### Get Metric Score Range

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/metrics/api.md

Retrieve the minimum and maximum possible values for a validation score. The output is a 2-tuple containing the minimum and maximum float values.

```php
public range() : Rubix\ML\Tuple{float, float}
```

```php
[$min, $max] = $metric->range()->list();

echo "min: $min, max: $max";
```

```text
min: 0.0, max: 1.0
```

--------------------------------

### Initialize Labeled Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/labeled.md

Create a new Labeled dataset instance by providing a 2D array of samples and a 1D array of corresponding labels. The 'verify' parameter can be set to false to skip data validation.

```php
use Rubix\ML\Datasets\Labeled;

$samples = [
    [0.1, 20, 'furry'],
    [2.0, -5, 'rough'],
    [0.01, 5, 'furry'],
];

$labels = ['not monster', 'monster', 'not monster'];

$dataset = new Labeled($samples, $labels);
```

--------------------------------

### Chain Multiple Transformers

Source: https://github.com/rubixml/ml/blob/master/docs/preprocessing.md

Apply multiple transformers sequentially to a dataset by fluently calling the apply() method. This example chains imputation, one-hot encoding, and normalization.

```php
use Rubix\ML\Transformers\HotDeckImputer;
use Rubix\ML\Transformers\OneHotEncoder;
use Rubix\ML\Transformers\MinMaxNormalizer;

$dataset->apply(new HotDeckImputer(5))
    ->apply(new OneHotEncoder())
    ->apply(new MinMaxNormalizer());
```

--------------------------------

### Instantiate Sigmoid Activation Function

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/activation-functions/sigmoid.md

Demonstrates how to create an instance of the Sigmoid activation function. This function has no configurable parameters.

```php
use Rubix\ML\NeuralNet\ActivationFunctions\Sigmoid;

$activationFunction = new Sigmoid();
```

--------------------------------

### Initialize Stochastic Optimizer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/optimizers/stochastic.md

Instantiate the Stochastic optimizer with a specified learning rate. This optimizer is suitable for basic gradient descent applications.

```php
use Rubix\ML\NeuralNet\Optimizers\Stochastic;

$optimizer = new Stochastic(0.01);
```

--------------------------------

### Initialize ImageResizer Transformer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/image-resizer.md

Instantiate the ImageResizer transformer with the desired width and height. The GD extension must be enabled.

```php
use Rubix\ML\Transformers\ImageResizer;

$transformer = new ImageResizer(28, 28);
```

--------------------------------

### Get Feature Importances from a Trained Ridge Regressor

Source: https://github.com/rubixml/ml/blob/master/docs/training.md

Train a Ridge regressor and then retrieve the importance scores for each feature. This is useful for feature selection and understanding model behavior.

```php
use Rubix\ML\Regressors\Ridge;

$estimator = new Ridge();

$estimator->train($dataset);

$importances = $estimator->featureImportances();

print_r($importances);
```

```sh
Array
(
    [0] => 0.04757
    [1] => 0.37948
    [2] => 0.53170
    [3] => 0.04123
)
```

--------------------------------

### Get Dropped Record Count

Source: https://github.com/rubixml/ml/blob/master/docs/extractors/deduplicator.md

Retrieve the total number of records that have been dropped by the Deduplicator due to being identified as duplicates. This method is useful for monitoring the effectiveness of the deduplication process.

```php
public function dropped() : int
```

--------------------------------

### Initialize KMostFrequent Strategy

Source: https://github.com/rubixml/ml/blob/master/docs/strategies/k-most-frequent.md

Instantiates the KMostFrequent strategy with a specified number of top classes to consider. This strategy is suitable for categorical data.

```php
use Rubix\ML\Strategies\KMostFrequent;

$strategy = new KMostFrequent(5);
```

--------------------------------

### Predicting with a Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/estimator.md

Use the `predict()` method to get predictions for unknown samples in a dataset. The return value is an array of predictions corresponding to the dataset's indexed order.

```php
public predict(Dataset $dataset) : array
```

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

print_r($predictions);
```

```text
Array
(
    [0] => married
    [1] => divorced
    [2] => divorced
    [3] => married
)
```

--------------------------------

### Update Stateful Transformer Fitting

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/api.md

Elastic transformers can adapt to new data using the update() method after being fitted. This example updates a ZScaleStandardizer with new data folds.

```php
use Rubix\ML\Transformers\ZScaleStandardizer;

$transformer = new ZScaleStandardizer();

$folds = $dataset->fold(3);

$transformer->fit($folds[0]);

$transformer->update($folds[1]);

$transformer->update($folds[2]);
```

--------------------------------

### Instantiate Completeness Metric

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/metrics/completeness.md

Instantiates the Completeness metric. This metric does not require any parameters.

```php
use Rubix\ML\CrossValidation\Metrics\Completeness;

$metric = new Completeness();
```

--------------------------------

### Fit a Stateful Transformer to a Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/api.md

Stateful transformers require fitting before transformation. The fit() method takes a Dataset and pre-computes necessary information. This example uses OneHotEncoder.

```php
use Rubix\ML\Transformers\OneHotEncoder;

$transformer = new OneHotEncoder();

$transformer->fit($dataset);

var_dump($transformer->fitted());
```

--------------------------------

### Instantiate Native Serializer

Source: https://github.com/rubixml/ml/blob/master/docs/serializers/native.md

This snippet shows how to create an instance of the Native serializer. No parameters are required for initialization.

```php
use Rubix\ML\Serializers\Native;

$serializer = new Native();
```

--------------------------------

### Instantiate a Labeled Dataset

Source: https://github.com/rubixml/ml/blob/master/docs/datasets/api.md

Create a new Labeled dataset object by passing samples and their corresponding labels to the constructor. Ensure labels are provided as a separate array.

```php
use Rubix\ML\Datasets\Labeled;

$samples = [
    [0.1, 20, 'furry'],
    [2.0, -5, 'rough'],
];

$labels = ['not monster', 'monster'];

$dataset = new Labeled($samples, $labels);
```

--------------------------------

### Making Predictions with an Estimator

Source: https://github.com/rubixml/ml/blob/master/docs/inference.md

Use the `predict()` method on a trained estimator to get predictions for a dataset. Ensure the inference dataset has the same number and order of feature columns as the training data.

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

print_r($predictions);
```

```plaintext
Array
(
    [0] => cat
    [1] => dog
    [2] => frog
)
```

--------------------------------

### Instantiate SparseRandomProjector

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/sparse-random-projector.md

Instantiate the SparseRandomProjector transformer with a specified number of target dimensions and automatic sparsity selection.

```php
use Rubix\ML\Transformers\SparseRandomProjector;

$transformer = new SparseRandomProjector(30, null);
```

--------------------------------

### Instantiate Gradient Boost Regressor

Source: https://github.com/rubixml/ml/blob/master/docs/regressors/gradient-boost.md

Instantiates a GradientBoost regressor with specified parameters. This includes the base learner, learning rate, subsampling ratio, maximum epochs, minimum change threshold, early stopping window, hold-out ratio for validation, and the evaluation metric.

```php
use Rubix\ML\Regressors\GradientBoost;
use Rubix\ML\Regressors\RegressionTree;
use Rubix\ML\CrossValidation\Metrics\SMAPE;

$estimator = new GradientBoost(new RegressionTree(3), 0.1, 0.8, 1000, 1e-4, 10, 0.1, new SMAPE());
```

--------------------------------

### Instantiate RMSProp Optimizer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/optimizers/rms-prop.md

Instantiate the RMSProp optimizer with a specified learning rate and decay rate.

```php
use Rubix\ML\NeuralNet\Optimizers\RMSProp;

$optimizer = new RMSProp(0.01, 0.1);
```

--------------------------------

### Instantiate LambdaFunction Transformer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/lambda-function.md

This snippet shows how to create a LambdaFunction transformer. It defines a callback function that adds a transformed value to the sample and then instantiates the transformer with the callback and an example context.

```php
use Rubix\ML\Transformers\LambdaFunction;

$callback = function (&$sample, $offset, $context) {
	$sample[] = log1p($sample[3]);
};

$transformer = new LambdaFunction($callback, 'example context');
```

--------------------------------

### Instantiate Min Max Normalizer

Source: https://github.com/rubixml/ml/blob/master/docs/transformers/min-max-normalizer.md

Instantiate the MinMaxNormalizer with custom minimum and maximum values for the transformation range.

```php
use Rubix\ML\Transformers\MinMaxNormalizer;

$transformer = new MinMaxNormalizer(-5.0, 5.0);
```

--------------------------------

### Initialize Gaussian Mixture Model

Source: https://github.com/rubixml/ml/blob/master/docs/clusterers/gaussian-mixture.md

Instantiate a Gaussian Mixture model with specified parameters including the number of clusters, smoothing, epochs, minimum change, and a seeder.

```php
use Rubix\ML\Clusterers\GaussianMixture;
use Rubix\ML\Clusterers\Seeders\KMC2;

$estimator = new GaussianMixture(5, 1e-6, 100, 1e-4, new KMC2(50));
```

--------------------------------

### Initialize K-Means Estimator

Source: https://github.com/rubixml/ml/blob/master/docs/clusterers/k-means.md

Instantiate the KMeans estimator with custom parameters for the number of clusters, batch size, epochs, minimum change threshold, early stopping window, distance kernel, and seeder.

```php
use Rubix\ML\Clusterers\KMeans;
use Rubix\ML\Kernels\Distance\Euclidean;
use Rubix\ML\Clusterers\Seeders\PlusPlus;

$estimator = new KMeans(3, 128, 300, 10.0, 10, new Euclidean(), new PlusPlus());
```

--------------------------------

### Instantiate Concatenator with Multiple CSV Iterators

Source: https://github.com/rubixml/ml/blob/master/docs/extractors/concatenator.md

Instantiates the Concatenator extractor by providing an array of CSV iterators. This setup is ideal for merging data from several CSV files into one.

```php
use Rubix\ML\Extractors\Concatenator;
use Rubix\ML\Extractors\CSV;

$extractor = new Concatenator([
    new CSV('dataset1.csv'),
    new CSV('dataset2.csv'),
    new CSV('dataset3.csv'),
]);
```

--------------------------------

### Get Joint Probability Estimates

Source: https://github.com/rubixml/ml/blob/master/docs/probabilistic.md

Use the `proba()` method to retrieve joint probability estimates from a dataset. This method is useful for understanding the estimator's certainty about each possible class or cluster.

```php
public proba(Dataset $dataset) : array
```

```php
$probabilities = $estimator->proba($dataset);

print_r($probabilities);
```

```php
Array
(
    [0] => Array
        (
            [monster] => 0.6
            [not monster] => 0.4
        )

    [1] => Array
        (
            [monster] => 0.5
            [not monster] => 0.5
        )

    [2] => Array
        (
            [monster] => 0.2
            [not monster] => 0.8
        )
)

```

--------------------------------

### Instantiate Normal Initializer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/initializers/normal.md

Instantiate the Normal initializer with a custom standard deviation for the Gaussian distribution.

```php
use Rubix\ML\NeuralNet\Initializers\Normal;

$initializer = new Normal(0.1);
```

--------------------------------

### Getting Anomaly Scores with a Scoring Estimator

Source: https://github.com/rubixml/ml/blob/master/docs/inference.md

Anomaly detectors implementing the `Scoring` interface can use the `score()` method to output anomaly scores for samples. Higher scores indicate greater abnormality.

```php
$scores = $estimator->score($dataset);

print_r($scores);
```

```plaintext
Array
(
    [0] => 0.35033
    [1] => 0.40992
    [2] => 1.68153
)
```

--------------------------------

### Initialize Swish Layer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/hidden-layers/swish.md

Instantiate a Swish activation layer with a Constant initializer for the beta parameter.

```php
use Rubix\ML\NeuralNet\Layers\Swish;
use Rubix\ML\NeuralNet\Initializers\Constant;

$layer = new Swish(new Constant(1.0));
```

--------------------------------

### Instantiate Bootstrap Aggregator

Source: https://github.com/rubixml/ml/blob/master/docs/bootstrap-aggregator.md

Instantiate the BootstrapAggregator with a base learner, number of estimators, and sample ratio. This is useful for creating an ensemble of models to reduce variance.

```php
use Rubix\ML\BootstrapAggregator;
use Rubix\ML\Regressors\RegressionTree;

$estimator = new BootstrapAggregator(new RegressionTree(10), 300, 0.2);
```

--------------------------------

### Instantiate KSkipNGram Tokenizer

Source: https://github.com/rubixml/ml/blob/master/docs/tokenizers/k-skip-n-gram.md

Instantiates the KSkipNGram tokenizer with minimum token length 2, maximum token length 3, and a skip value of 2.

```php
use Rubix\ML\Tokenizers\KSkipNGram;

$tokenizer = new KSkipNGram(2, 3, 2);
```

--------------------------------

### Example Confusion Matrix JSON Output

Source: https://github.com/rubixml/ml/blob/master/docs/cross-validation/reports/confusion-matrix.md

A sample JSON structure representing a confusion matrix. Keys represent actual labels, and nested keys represent predicted labels with their corresponding counts.

```json
{
    "dog": {
        "dog": 12,
        "cat": 3,
        "turtle": 0
    },
    "cat": {
        "dog": 2,
        "cat": 9,
        "turtle": 1
    },
    "turtle": {
        "dog": 1,
        "cat": 0,
        "turtle": 11
    }
}
```

--------------------------------

### Instantiate a Dense Layer

Source: https://github.com/rubixml/ml/blob/master/docs/neural-network/hidden-layers/dense.md

Instantiates a Dense layer with a specified number of neurons, L2 penalty, bias inclusion, and weight/bias initializers. Ensure the necessary classes are imported before use.

```php
use Rubix\ML\NeuralNet\Layers\Dense;
use Rubix\ML\NeuralNet\Initializers\He;
use Rubix\ML\NeuralNet\Initializers\Constant;

$layer = new Dense(100, 1e-4, true, new He(), new Constant(0.0));
```

--------------------------------

### Export Neural Network Architecture to Image

Source: https://github.com/rubixml/ml/blob/master/docs/classifiers/multilayer-perceptron.md

Exports the neural network architecture as a Graphviz "dot" encoding and then converts it to a PNG image saved to 'network.png'. Ensure Graphviz is installed and accessible.

```php
use Rubix\ML\Helpers\Graphviz;
use Rubix\ML\Persisters\Filesystem;

$dot = $estimator->exportGraphviz();

Graphviz::dotToImage($dot)->saveTo(new Filesystem('network.png'));
```

--------------------------------

### Instantiate Wild Guess Strategy

Source: https://github.com/rubixml/ml/blob/master/docs/strategies/wild-guess.md

Instantiates the WildGuess strategy. This strategy does not require any parameters.

```php
use Rubix\ML\Strategies\WildGuess;

$strategy = new WildGuess();
```