PHP改进计算字符串相似度的函数similar_text()、levenshtein()

similar_text()中文汉字版

复制代码代码如下:

<?php
//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;
}

levenshtein()中文汉字版

复制代码代码如下:

<?php
//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);
}

最长公共子序列LCS()

复制代码代码如下:

<?php
//最长公共子序列英文版
function LCS_en($str_1, $str_2) {
$len_1 = strlen($str_1);
$len_2 = strlen($str_2);
$len = $len_1 > $len_2 ? $len_1 : $len_2;
$dp = array();
for ($i = 0; $i <= $len; $i++) {
$dp[$i] = array();
$dp[$i][0] = 0;
$dp[0][$i] = 0;
}
for ($i = 1; $i <= $len_1; $i++) {
for ($j = 1; $j <= $len_2; $j++) {
if ($str_1[$i – 1] == $str_2[$j – 1]) {
$dp[$i][$j] = $dp[$i – 1][$j – 1] + 1;
} else {
$dp[$i][$j] = $dp[$i – 1][$j] > $dp[$i][$j – 1] ? $dp[$i – 1][$j] : $dp[$i][$j – 1];
}
}
}
return $dp[$len_1][$len_2];
}
//拆分字符串
function mbStringToArray($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 $arrayResult;
}
//最长公共子序列中文版
function LCS_cn($str1, $str2, $encoding = 'UTF-8') {
$mb_len1 = mb_strlen($str1, $encoding);
$mb_len2 = mb_strlen($str2, $encoding);
$mb_str1 = mbStringToArray($str1, $encoding);
$mb_str2 = mbStringToArray($str2, $encoding);
$len = $mb_len1 > $mb_len2 ? $mb_len1 : $mb_len2;
$dp = array();
for ($i = 0; $i <= $len; $i++) {
$dp[$i] = array();
$dp[$i][0] = 0;
$dp[0][$i] = 0;
}
for ($i = 1; $i <= $mb_len1; $i++) {
for ($j = 1; $j <= $mb_len2; $j++) {
if ($mb_str1[$i – 1] == $mb_str2[$j – 1]) {
$dp[$i][$j] = $dp[$i – 1][$j – 1] + 1;
} else {
$dp[$i][$j] = $dp[$i – 1][$j] > $dp[$i][$j – 1] ? $dp[$i – 1][$j] : $dp[$i][$j – 1];
}
}
}
return $dp[$mb_len1][$mb_len2];
}