php-ml/docs/machine-learning/classification/svc.md
Yuji Uchiyama ec091b5ea3 Support probability estimation in SVC (#218)
* Add test for svm model with probability estimation

* Extract buildPredictCommand method

* Fix test to use PHP_EOL

* Add predictProbability method (not completed)

* Add test for DataTransformer::predictions

* Fix SVM to use PHP_EOL

* Support probability estimation in SVM

* Add documentation

* Add InvalidOperationException class

* Throw InvalidOperationException before executing libsvm if probability estimation is not supported
2018-02-06 20:39:25 +01:00

2.7 KiB
Raw Blame History

Support Vector Classification

Classifier implementing Support Vector Machine based on libsvm.

Constructor Parameters

  • $kernel (int) - kernel type to be used in the algorithm (default Kernel::LINEAR)
  • $cost (float) - parameter C of C-SVC (default 1.0)
  • $degree (int) - degree of the Kernel::POLYNOMIAL function (default 3)
  • $gamma (float) - kernel coefficient for Kernel::RBF, Kernel::POLYNOMIAL and Kernel::SIGMOID. If gamma is null then 1/features will be used instead.
  • $coef0 (float) - independent term in kernel function. It is only significant in Kernel::POLYNOMIAL and Kernel::SIGMOID (default 0.0)
  • $tolerance (float) - tolerance of termination criterion (default 0.001)
  • $cacheSize (int) - cache memory size in MB (default 100)
  • $shrinking (bool) - whether to use the shrinking heuristics (default true)
  • $probabilityEstimates (bool) - whether to enable probability estimates (default false)
$classifier = new SVC(Kernel::LINEAR, $cost = 1000);
$classifier = new SVC(Kernel::RBF, $cost = 1000, $degree = 3, $gamma = 6);

Train

To train a classifier simply provide train samples and labels (as array). Example:

use Phpml\Classification\SVC;
use Phpml\SupportVectorMachine\Kernel;

$samples = [[1, 3], [1, 4], [2, 4], [3, 1], [4, 1], [4, 2]];
$labels = ['a', 'a', 'a', 'b', 'b', 'b'];

$classifier = new SVC(Kernel::LINEAR, $cost = 1000);
$classifier->train($samples, $labels);

You can train the classifier using multiple data sets, predictions will be based on all the training data.

Predict

To predict sample label use predict method. You can provide one sample or array of samples:

$classifier->predict([3, 2]);
// return 'b'

$classifier->predict([[3, 2], [1, 5]]);
// return ['b', 'a']

Probability estimation

To predict probabilities you must build a classifier with $probabilityEstimates set to true. Example:

use Phpml\Classification\SVC;
use Phpml\SupportVectorMachine\Kernel;

$samples = [[1, 3], [1, 4], [2, 4], [3, 1], [4, 1], [4, 2]];
$labels = ['a', 'a', 'a', 'b', 'b', 'b'];

$classifier = new SVC(
    Kernel::LINEAR, // $kernel
    1.0,            // $cost
    3,              // $degree
    null,           // $gamma
    0.0,            // $coef0
    0.001,          // $tolerance
    100,            // $cacheSize
    true,           // $shrinking
    true            // $probabilityEstimates, set to true
);

$classifier->train($samples, $labels);

Then use predictProbability method instead of predict:

$classifier->predictProbability([3, 2]);
// return ['a' => 0.349833, 'b' => 0.650167]

$classifier->predictProbability([[3, 2], [1, 5]]);
// return [
//   ['a' => 0.349833, 'b' => 0.650167],
//   ['a' => 0.922664, 'b' => 0.0773364],
// ]