PHP调用scws分词计算tfidf 比较文本内容相似度

<?php
Class ContentSim {
    
	var $analysis;
	var $word_limit = 20;
    public function __construct(){
		$this->analysis = scws_new();
		$this->analysis->set_charset('utf8');
    }

	public function get_tfidf($txt){
		$word_ret = array();
		$this->analysis->send_text($txt);
		$total = 0;
		while($result = $this->analysis->get_result()){
			$total++;
			foreach($result as $v){
				if(mb_strlen($v['word'],"UTF-8")<=1) continue;
				if(preg_match("/[^\w\s]+/u",$v['word'])){
					continue;
				}
				if( array_key_exists($v['word'] , $word_ret) ){
					$v["cnt"]=intval($word_ret[$v["word"]]["cnt"])+1;
				}
				else{
					$v["cnt"] = 1;
				}
				$word_ret[$v["word"]] = $v;
			}
		}
		$sort = array();
		foreach ($word_ret as $key => $row) {
			$sort[$key]  = round(($row['cnt']/$total) * $row['idf'],2);
		}
		unset($word_ret);
		arsort($sort);
		return $sort;
	}

	public function combine_word($arr1,$arr2){
		$word_array = array();
        foreach($arr1 as $k=>$v){
			$word_array[$k] = array($v,0);
        }
        foreach($arr2 as $k=>$v){
			if( !array_key_exists($k , $word_array) ){
				$word_array[$k] = array(0,$v);
			}else{
				$word_array[$k][1] = $v;
            }
        }
		return $word_array;
	}

    public function sim_value($arr1,$arr2) {
		$words = $this->combine_word($arr1,$arr2);
		unset($arr1);unset($arr2);
        $c = 0;$s1 = $s2 = 0;
		foreach($words as $row){
			$c += floatval($row[0]) * floatval($row[1]);
            $s1  += pow(floatval($row[0]),2);
            $s2  += pow(floatval($row[1]),2);
		}
        $f = $c / (sqrt($s1) * sqrt($s2));
        return $f;
    }
	public function close_analysis(){
		$this->analysis->close();
	}
    
}
$txt1 = '';
$txt2 = '';
$ContentSim = new ContentSim ( );

$txt_ret1 = $ContentSim->get_tfidf($text1);
$txt_ret2 = $ContentSim->get_tfidf($text2);


$result = $ContentSim->sim_value($txt_ret1,$txt_ret2);

?>