1<span style="float:right;"><a href="https://github.com/RubixML/ML/blob/master/src/GridSearch.php">[source]</a></span>
2
3# Grid Search
4Grid Search is an algorithm that optimizes hyper-parameter selection. From the user's perspective, the process of training and predicting is the same, however, under the hood Grid Search trains a model for each combination of possible parameters and the best model is selected as the base estimator.
5
6**Interfaces:** [Wrapper](wrapper.md), [Estimator](estimator.md), [Learner](learner.md), [Parallel](parallel.md), [Persistable](persistable.md), [Verbose](verbose.md)
7
8**Data Type Compatibility:** Depends on base learner
9
10## Parameters
11| # | Name | Default | Type | Description |
12|---|---|---|---|---|
13| 1 | base | | string | The class name of the base learner. |
14| 2 | params | | array | An array of [n-tuples](faq.md#what-is-a-tuple) containing the possible values for each of the base learner's constructor parameters. |
15| 3 | metric | null | Metric | The validation metric used to score each set of hyper-parameters. If null, automatically selects a default metric based on estimator type. |
16| 4 | validator | KFold | Validator | The validator used to test and score each trained model. |
17
18## Example
19```php
20use Rubix\ML\GridSearch;
21use Rubix\ML\Classifiers\KNearestNeighbors;
22use Rubix\ML\Kernels\Distance\Euclidean;
23use Rubix\ML\Kernels\Distance\Manhattan;
24use Rubix\ML\CrossValidation\Metrics\FBeta;
25use Rubix\ML\CrossValidation\KFold;
26
27$params = [
28	[1, 3, 5, 10], [true, false], [new Euclidean(), new Manhattan()],
29];
30
31$estimator = new GridSearch(KNearestNeighbors::class, $params, new FBeta(), new KFold(5));
32```
33
34## Additional Methods
35Return an array containing the validation scores and hyper-parameters under test for each combination resulting from the last search:
36```php
37public results() : ?array
38```
39
40Return an array containing the best parameters from the last search:
41```php
42public best() : ?array
43```
44
45```php
46
47var_dump($estimator->best());
48```
49
50```sh
51array(3) {
52  [0]=> int(3)
53  [1]=> bool(true)
54  [2]=> object(Rubix\ML\Kernels\Distance\Manhattan) {}
55}
56```