php simple Chinese word segmentation system

php tutorial simple Chinese word segmentation system

Structure: first hash table, trie index tree node Advantages: word segmentation, without the need to predict the length of the query word, word by word matching along the tree.
Disadvantages: the construction and maintenance more complicated, the word branches, wasting a certain amount of space
* @ version 0.1
* @todo Constructs a general dictionary algorithm and writes a simple word segmentation
* @author shjuto@gmail.com
* trie dictionary tree
*
* /

class trie
{
private

function __construct ()
{
$ trie = array ('children' => array (), 'isword' => false);
}

/ **
* Add words to the dictionary
*
* @param string $ key
* /
function & setword ($ word = '')
{
$ trienode = & $ this-> trie;
for ($ i = 0; $ i <strlen ($ word); $ i ++)
{
$ character = $ word [$ i];
if (! isset ($ trienode ['children'] [$ character]))
{
$ trienode ['children'] [$ character] = array ('isword' => false);
}
if ($ i == strlen ($ word) -1)
{
$ trienode ['children'] [$ character] = array ('isword' => true);
}
$ trienode = & $ trienode ['children'] [$ character];
}
}

/ **
* To determine whether the dictionary word
*
* @param string $ word
@return bool true / false
* /
function & isword ($ word)
{
$ trienode = & $ this-> trie;
for ($ i = 0; $ i <strlen ($ word); $ i ++)
{
$ character = $ word [$ i];
if (! isset ($ trienode ['children'] [$ character]))
{
return false;
}
else
{
// judge the end of the word
if ($ i == (strlen ($ word) -1) && $ trienode ['children'] [$ character] ['isword'] == true)
{
return true;
}
elseif ($ i == (strlen ($ word) -1) && $ trienode ['children'] [$ character] ['isword'] == false)
{
return false;
}
$ trienode = & $ trienode ['children'] [$ character];
}
}
}

/ **
* The location where the text $ text appears
*
* @param string $ text
* @return array array ('position' => $ position, 'word' => $ word);
* /
function search ($ text = "")
{
$ textlen = strlen ($ text);
$ trienode = $ tree = $ this-> trie;
$ find = array ();
$ wordrootposition = 0; // stem location
$ prenode = false; / / backtracking parameters, when the dictionary ab, in the string aab, $ i need to backtrack back once
$ word = '';
for ($ i = 0; $ i <$ textlen; $ i ++)
{

if (isset ($ trienode ['children'] [$ text [$ i]]))
{
$ word = $ word. $ text [$ i];
$ trienode = $ trienode ['children'] [$ text [$ i]];
if ($ prenode == false)
{
$ wordrootposition = $ i;
}
$ prenode = true;
if ($ trienode ['isword'])
{
$ find [] = array ('position' => $ wordrootposition, 'word' => $ word);
}
}
else
{
$ trienode = $ tree;
$ word = '';
if ($ prenode)
{
$ i = $ i -1;
$ prenode = false;
}
}
}
return $ find;
}
}
$ trie = new trie ();
$ trie-> setword ('China');
$ trie-> setword ('Chinese');
$ trie-> setword ('great');
$ trie-> setword ('army');
$ trie-> setword ('Chinese people');
$ trie-> setword ('People's Liberation Army');
$ trie-> setword ('People's Liberation Army');
$ trie-> setword ('liberation');
$ words = $ trie-> search ('The great liberation of the Chinese People's Liberation Army of China is a very great army');
foreach ($ words as $ word)
{
echo 'position:'. $ word ['position'] .'- '(strlen ($ word [' word ']) + $ word [' position ']);
echo 'word:'. $ word ['word']. "n";
}