字符串相似度检测

字符串相似度检测

1、similar_text()
2、levenshtein()
$sum = mb_strlen($s1)+mb_strlen($s2);
$dist = levenshtein($s1,$s2,1,2,1);
$ratio = ($sum-$dist)/$sum;

3、改进版similar_text()中文汉字版
//拆分字符串
function split_str($str) {
preg_match_all("/./u", $str, $arr);
return $arr[0];
}//相似度检测
function similar_text_cn($str1, $str2) {
$arr_1 = array_unique(split_str($str1));
$arr_2 = array_unique(split_str($str2));
$similarity = count($arr_2) - count(array_diff($arr_2, $arr_1));return $similarity;
}

4、levenshtein()中文汉字版

//拆分字符串
function mbStringToArray($string, $encoding = 'UTF-8') {
$arrayResult = array();
while ($iLen = mb_strlen($string, $encoding)) {
array_push($arrayResult, mb_substr($string, 0, 1, $encoding));
$string = mb_substr($string, 1, $iLen, $encoding);
}
return $arrayResult;
}
//编辑距离
function levenshtein_cn($str1, $str2, $costReplace = 1, $encoding = 'UTF-8') {
$count_same_letter = 0;
$d = array();
$mb_len1 = mb_strlen($str1, $encoding);
$mb_len2 = mb_strlen($str2, $encoding);
$mb_str1 = mbStringToArray($str1, $encoding);
$mb_str2 = mbStringToArray($str2, $encoding);
for ($i1 = 0; $i1 <= $mb_len1; $i1++) {
$d[$i1] = array();
$d[$i1][0] = $i1;
}
for ($i2 = 0; $i2 <= $mb_len2; $i2++) {
$d[0][$i2] = $i2;
}
for ($i1 = 1; $i1 <= $mb_len1; $i1++) {
for ($i2 = 1; $i2 <= $mb_len2; $i2++) {
// $cost = ($str1[$i1 - 1] == $str2[$i2 - 1]) ? 0 : 1;
if ($mb_str1[$i1 - 1] === $mb_str2[$i2 - 1]) {
$cost = 0;
$count_same_letter++;
} else {
$cost = $costReplace; //替换
}
$d[$i1][$i2] = min($d[$i1 - 1][$i2] + 1, //插入
$d[$i1][$i2 - 1] + 1, //删除
$d[$i1 - 1][$i2 - 1] + $cost);
}
}
return $d[$mb_len1][$mb_len2];
//return array('distance' => $d[$mb_len1][$mb_len2], 'count_same_letter' => $count_same_letter);
}
//获取两者中最长的字符串长度
$len=mb_strlen($str1)>=mb_strlen($str2)?mb_strlen($str1):mb_strlen($str2);
//计算相似度,并输出
echo 1-levenshtein_cn($str1, $str2)/$len;