VectorSpaceModel extension for php-nlp-tools

VectorSpaceModel.php

<?php

namespace App\Utilities\Search;

use App\Library\InverseDocumentFrequency;
use App\Utilities\Search\TfIdfFeatureFactory;
use NlpTools\Documents\TrainingSet;


/**
 * Vector Space Model is a Class for calculating Relevance ranking by comparing the deviation of angles between each 
 * document vector and the original query vector.
 *
 * You can use current implementation of php-nlp-tools' cosine similarity but it was made to return an Exception in case of
 * 0 vector product instead of returning 0.
 *
 * https://en.wikipedia.org/wiki/Vector_space_model#Example:_tf-idf_weights
 *
 * @author Jericko Tejido <[email protected]>
 */

class VectorSpaceModel
{


    protected $tset;


    public function __construct(TrainingSet $tset) 
    {
        $this->tset = $tset;
    }


    /**
     * Returns Score ranking per Documents added by ascending order.
     *
     * @param  string $term
     * @return array
     */
    public function getScores(TrainingSet $qset)
    {

        $score = array();
        $term_tfidf = array();
        $doc_tfidf = array();

        $docs_idf = new InverseDocumentFrequency($this->tset);
        $ff = new TfIdfFeatureFactory(
            $docs_idf,
            array(
                function ($c, $d) {
                    return $d->getDocumentData();
                }
            )
        );

        for($i = 0; $i < count($this->tset); $i++){
            ;
            $id = $this->tset->offsetGet($i)->getClass();
            $query = $ff->getFeatureArray("", $qset->offsetGet(0));
            $documents = $ff->getFeatureArray("", $this->tset->offsetGet($i));
            $score[$id] = $this->getSimilarity($query, $documents);
        }

        arsort($score);
        return $score;

    }

    /**
     * returns Euclidean norm
     *
     * @param array $vector
     * @return mixed
     */
    private function norm(array $vector) {
        return sqrt($this->dotProduct($vector, $vector));
    }

    /**
     * returns Dot product
     *
     * @param array $a
     * @param array $b
     * @return mixed
     */
    private function dotProduct(array $a, array $b) {
        $dotProduct = 0;
        $keysA = array_keys(array_filter($a));
        $keysB = array_keys(array_filter($b));
        $uniqueKeys = array_unique(array_merge($keysA, $keysB));
        foreach ($uniqueKeys as $key) {
            if (!empty($a[$key]) && !empty($b[$key]))
                $dotProduct += ($a[$key] * $b[$key]);
        }
        return $dotProduct;
    }

    /**
     * returns cos(theta) between two non-normalised vectors
     * sim(a, b) = (a・b) / (||a|| * ||b||)
     *
     * @param array $a
     * @param array $b
     * @return mixed
     */
    private function getSimilarity(array $a, array $b) {
        $normA = $this->norm($a);
        $normB = $this->norm($b);
        return (($normA * $normB) != 0)
               ? $this->dotProduct($a, $b) / ($normA * $normB)
               : 0;
    }

}