PHP Jaccard 유사성 지수 구현

PHP Jaccard 유사성 지수 구현
PHP 자카드 유사도

1 함수[ | ]

function jaccard_index($set1, $set2) {
	$union_cnt = count(array_unique(array_merge($set1, $set2)));
	if($union_cnt < 1) return 1;
	$intersect_cnt = count(array_intersect($set1, $set2));
	return $intersect_cnt / $union_cnt;
}

2 예제[ | ]

<?php
function jaccard_index($set1, $set2) {
	$union_cnt = count(array_unique(array_merge($set1, $set2)));
	if($union_cnt < 1)return 1;
	$intersect_cnt = count(array_intersect($set1, $set2));
	return $intersect_cnt / $union_cnt;
}

$sets = array();
$sets['A'] = array(3,5,6,7,8);
$sets['B'] = array(1,2,3,4,5,6,8,9,10);
$sets['C'] = array(2,3,9);
$sets['D'] = array(1,2,3,4,5,6,8);

$cnt = count($sets);
echo "<table>";
echo "<th>구분</th>";
foreach(array_keys($sets) as $i) {
	echo "<th>$i</th>";
}
foreach($sets as $i => $set1) {
	echo "<tr>";
	echo "<th>$i</th>";
	foreach($sets as $j => $set2) {
		if($i == $j) {
			$j_index = '-';
		} else {
			$j_index = jaccard_index($set1, $set2);
			$j_index = number_format($j_index, 2);
		}
		echo "<td>$j_index</td>";
	}
	echo "</tr>";
}
echo "</table>";

http://zetawiki.com/ex/php/jaccard_index.php

구분	A	B	C	D
A	-	0.40	0.14	0.50
B	0.40	-	0.33	0.78
C	0.14	0.33	-	0.25
D	0.50	0.78	0.25	-

→ 집합 B와 D는 유사성 0.78 → 상당히 유사함

3 같이 보기[ | ]