<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> <html> <head> <meta name="generator" content="HTML Tidy, see www.w3.org"> <title>MeCab: Yet Another Part-of-Speech and Morphological Analyzer</title> <link type="text/css" rel="stylesheet" href="mecab.css"> </head> <body> <h1>MeCab: Yet Another Part-of-Speech and Morphological Analyzer</h1> <p>$Id: index.html 135 2007-06-10 12:28:13Z taku-ku $;</p> <h2>MeCab (ÏÂÉÛÉó)¤È¤Ï</h2> <p> MeCab¤Ï<a href="http://pine.kuee.kyoto-u.ac.jp/KU-NTT-WS-2005/"> µþÅÔÂç³Ø¾ðÊó³Ø¸¦µæ²Ê¡ÝÆüËÜÅÅ¿®ÅÅÏóô¼°²ñ¼Ò¥³¥ß¥å¥Ë¥±¡¼¥·¥ç¥ó²Ê³Ø´ðÁø¦µæ½ê ¶¦Æ±¸¦µæ¥æ¥Ë¥Ã¥È¥×¥í¥¸¥§¥¯¥È</a>¤òÄ̤¸¤Æ³«È¯¤µ¤ì¤¿¥ª¡¼¥×¥ó¥½¡¼¥¹ ·ÁÂÖÁDzòÀÏ¥¨¥ó¥¸¥ó¤Ç¤¹. ¸À¸ì, ¼½ñ,¥³¡¼¥Ñ¥¹¤Ë°Í¸¤·¤Ê¤¤ÈÆÍÑŪ¤ÊÀ߷פò ´ðËÜÊý¿Ë¤È¤·¤Æ¤¤¤Þ¤¹. ¥Ñ¥é¥á¡¼¥¿¤Î¿äÄê¤Ë Conditional Random Fields (<a href="http://www.cis.upenn.edu/~pereira/papers/crf.pdf">CRF</a>) ¤òÍÑ ¤¤¤Æ¤ª¤ê, <a href="http://chasen.naist.jp">ChaSen</a>¤¬ºÎÍѤ·¤Æ¤¤¤ë ±£¤ì¥Þ¥ë¥³¥Õ¥â¥Ç¥ë¤ËÈæ¤ÙÀǽ¤¬¸þ¾å¤·¤Æ¤¤¤Þ¤¹¡£¤Þ¤¿¡¢Ê¿¶ÑŪ¤Ë <a href="http://chasen.naist.jp">ChaSen</a>, <a href="http://www.kc.t.u-tokyo.ac.jp/nl-resource/juman.html">Juman</a>, <a href="http://kakasi.namazu.org">KAKASI</a>¤è¤ê¹â®¤ËÆ°ºî¤·¤Þ¤¹. ¤Á¤Ê¤ß¤ËÏÂÉÛÉó(¤á¤«¤Ö)¤Ï, ºî¼Ô¤Î¹¥Êª¤Ç¤¹. </p> </ul> <h2>Ìܼ¡</h2> <ul> <li><a href="#feature">ÆÃħ</a></li> <li><a href="#diff">Èæ³Ó</a></li> <li><a href="#news">¿·Ãå¾ðÊó</a></li> <li><a href="feature.html">³«È¯¤Þ¤Ç¤Î·Ð°Þ</a></li> <li><a href="#download">¥À¥¦¥ó¥í¡¼¥É</a></li> <li><a href="#install">¥¤¥ó¥¹¥È¡¼¥ë</a> <ul> <li><a href="#install-unix">Unix</a></li> <li><a href="#install-windows">Windows</a></li> </ul> </li> <li><a href="#usage-tools">»È¤¤Êý</a> <ul> <li><a href="#parse">¤È¤ê¤¢¤¨¤º²òÀϤ¹¤ë</a></li> <li><a href="#wakati">¤ï¤«¤Á½ñ¤¤ò¤¹¤ë</a></li> <li><a href="#format">½ÐÎÏ¥Õ¥©¡¼¥Þ¥Ã¥È¤ÎÊѹ¹</a> </ul> <li><a href="#usage-tools2">¹âÅ٤ʻȤ¤Êý</a> <ul> <li><a href="#charset">ʸ»ú¥³¡¼¥É¤ÎÊѹ¹</a></li> <li><a href="#nbest">N-Best ²ò¤Î½ÐÎÏ</a></li> <li><a href="dic.html">ñ¸ì¤ÎÄɲÃÊýË¡</a></li> <li><a href="format.html">½ÐÎÏ¥Õ¥©¡¼¥Þ¥Ã¥È¤Î¾ÜºÙÄêµÁ</a></li> <li><a href="posid.html">ÉÊ»ìID¤ÎÄêµÁ</a></li> <li><a href="partial.html">À©ÌóÉÕ¤²òÀÏ(Éôʬ²òÀÏ)</a></li> <li><a href="soft.html">¥½¥Õ¥È¤ï¤«¤Á½ñ¤</a></li> <li><a href="libmecab.html">C/C++ ¥é¥¤¥Ö¥é¥ê¥¤¥ó¥¿¥Õ¥§¡¼¥¹</a> </li> <li><a href="mecab.html">¤½¤Î¾¤Î¥³¥Þ¥ó¥É¥é¥¤¥ó¥ª¥×¥·¥ç¥ó</a></li> <li><a href="dic-detail.html">MeCab ¤Î¼½ñ¹½Â¤¤ÈÈÆÍѥƥ¥¹¥ÈÊÑ´¹¥Ä¡¼¥ë¤È¤·¤Æ¤ÎÍøÍÑ</a></li> <li><a href="unk.html">̤Ãθì½èÍý¤ÎºÆÄêµÁ</a></li> <li><a href="learn.html">¥ª¥ê¥¸¥Ê¥ë¼½ñ/¥³¡¼¥Ñ¥¹¤«¤é¤Î¥Ñ¥é¥á¡¼¥¿¿äÄê</a></li> <li><a href="bindings.html">¥¹¥¯¥ê¥×¥È¸À¸ì(perl/ruby/python/Java) ¥Ð¥¤¥ó¥Ç¥£¥ó¥°</a></li> </ul> <li><a href="#thanks">¼Õ¼</a></li> </ul> <h2><a name="feature">ÆÃħ</a></h2> <ul> <li>¼½ñ, ¥³¡¼¥Ñ¥¹¤Ë°Í¸¤·¤Ê¤¤ÈÆÍÑŪ¤ÊÀß·×</li> <li>¾ò·ïÉÕ¤³ÎΨ¾ì(<a href="http://www.cis.upenn.edu/~pereira/papers/crf.pdf">CRF</a>)¤Ë´ð¤Å¤¯¹â¤¤²òÀÏÀºÅÙ <li><a href="http://chasen.naist.jp">ChaSen</a> ¤ä <a href="http://kakasi.namazu.org">KAKASI</a> ¤ËÈæ¤Ù¹â®</li> <li>¼½ñ°ú¤¥¢¥ë¥´¥ê¥º¥à/¥Ç¡¼¥¿¹½Â¤¤Ë, ¹â®¤Ê TRIE ¹½Â¤¤Ç¤¢¤ë <a href="http://cl.naist.jp/~taku-ku/software/darts">Double-Array</a>¤òºÎÍÑ. <li>ºÆÆþ²Äǽ¤Ê¥é¥¤¥Ö¥é¥ê</li> <li>³Æ¼ï¥¹¥¯¥ê¥×¥È¸À¸ì¥Ð¥¤¥ó¥Ç¥£¥ó¥°(perl/ruby/python/java/C#)</li> </ul> <h2><a name="diff">Èæ³Ó</a></h2> <table> <tr class="even"> <td align="center"></td> <td align="center"><b>MeCab</b></td> <td align="center"><a href= "http://chasen.naist.jp/">ChaSen</a></td> <td align="center"><a href="http://pine.kuee.kyoto-u.ac.jp/nl-resource/juman.html">JUMAN</a></td> <td align="center"><a href="http://kakasi.namazu.org">KAKASI</a></td> </tr> <tr class="odd"> <td align="center">²òÀÏ¥â¥Ç¥ë</td> <td align="center">bi-gram ¥Þ¥ë¥³¥Õ¥â¥Ç¥ë</td> <td align="center">²ÄÊÑĹ¥Þ¥ë¥³¥Õ¥â¥Ç¥ë</td> <td align="center">bi-gram ¥Þ¥ë¥³¥Õ¥â¥Ç¥ë</td> <td align="center">ºÇĹ°ìÃ×</td> </tr> <tr class="even"> <td align="center">¥³¥¹¥È¿äÄê</td> <td align="center">¥³¡¼¥Ñ¥¹¤«¤é³Ø½¬</td> <td align="center">¥³¡¼¥Ñ¥¹¤«¤é³Ø½¬</td> <td align="center">¿Í¼ê</td> <td align="center">¥³¥¹¥È¤È¤¤¤¦³µÇ°Ìµ¤·</td> </tr> <tr class="odd"> <td align="center">³Ø½¬¥â¥Ç¥ë</td> <td align="center"><a href="http://www.cis.upenn.edu/~pereira/papers/crf.pdf">CRF</a> (¼±ÊÌ¥â¥Ç¥ë)</td> <td align="center">HMM (À¸À®¥â¥Ç¥ë)</td> <td align="center"></td> <td align="center"></td> </tr> <tr class="even"> <td align="center">¼½ñ°ú¤¥¢¥ë¥´¥ê¥º¥à</td> <td align="center">Double Array</td> <td align="center">Double Array</td> <td align="center">¥Ñ¥È¥ê¥·¥¢ÌÚ</td> <td align="center">Hash?</td> </tr> <tr class="odd"> <td align="center">²òõº÷¥¢¥ë¥´¥ê¥º¥à</td> <td align="center">Viterbi</td> <td align="center">Viterbi</td> <td align="center">Viterbi</td> <td align="center">·èÄêŪ?</td> </tr> <tr class="even"> <td align="center">Ï¢ÀÜɽ¤Î¼ÂÁõ</td> <td align="center">2¼¡¸µ Table</td> <td align="center">¥ª¡¼¥È¥Þ¥È¥ó</td> <td align="center">2¼¡¸µ Table?</td> <td align="center">Ï¢ÀÜɽ̵¤·?</td> </tr> <tr class="odd"> <td align="center">ÉÊ»ì¤Î³¬ÁØ</td> <td align="center">̵À©¸Â¿³¬ÁØÉÊ»ì</td> <td align="center">̵À©¸Â¿³¬ÁØÉÊ»ì</td> <td align="center">2Ãʳ¬¸ÇÄê</td> <td align="center">ÉÊ»ì¤È¤¤¤¦³µÇ°Ìµ¤·?</td> </tr> <tr class="even"> <td align="center">̤Ãθì½èÍý</td> <td align="center">»ú¼ï (Æ°ºîÄêµÁ¤òÊѹ¹²Äǽ)</td> <td align="center">»ú¼ï (Êѹ¹ÉÔ²Äǽ)</td> <td align="center">»ú¼ï (Êѹ¹ÉÔ²Äǽ)</td> <td align="center"></td> </tr> <tr class="odd"> <td align="center">À©Ìó¤Ä¤²òÀÏ</td> <td align="center">²Äǽ</td> <td align="center">2.4.0¤Ç²Äǽ</td> <td align="center">ÉÔ²Äǽ</td> <td align="center">ÉÔ²Äǽ</td> </tr> <tr class="even"> <td align="center">N-best²ò</td> <td align="center">²Äǽ</td> <td align="center">ÉÔ²Äǽ</td> <td align="center">ÉÔ²Äǽ</td> <td align="center">ÉÔ²Äǽ</td> </tr> </table> <p>MeCab ¤Ë»ê¤ë¤Þ¤Ç¤Î·ÁÂÖÁDzòÀϴﳫȯ¤ÎÎò»ËÅù¤Ï<a href="feature.html">¤³¤Á¤é</a>¤ò¤´Í÷¤¯¤À¤µ¤¤</li> <h2>¥á¡¼¥ê¥ó¥°¥ê¥¹¥È</h2> <ul> <li><a href="http://lists.sourceforge.jp/mailman/listinfo/mecab-users"> °ìÈ̥桼¥¶¸þ¤±¥á¡¼¥ê¥ó¥°¥ê¥¹¥È</a> <li><a href="http://lists.sourceforge.jp/mailman/listinfo/mecab-devel"> ³«È¯¼Ô¸þ¤±¥á¡¼¥ê¥ó¥°¥ê¥¹¥È</a> </ul> <h2><a name="news">¿·Ãå¾ðÊó</a></h2> <ul> <li><strong>2007-06-10</strong> MeCab 0.96<br> <ul> <li>¥Ð¥Ã¥Õ¥¡¥ª¡¼¥Ð¥Õ¥í¡¼¤Î¥Ð¥°¤ò½¤Àµ <li>¾ï¤ËPOS-ID¤òºîÀ®¤¹¤ë¤è¤¦¤Ë¤·¤¿ (-p ¥ª¥×¥·¥ç¥ó¤ÎÇÑ»ß) <li>¥æ¡¼¥¶¼½ñ¤Î¥Ç¥ê¥ß¥¿¤ò : ¤«¤é , (CSV) ¤ËÊѹ¹ (WindowsÂкö) <li>charset¤ÎȽÄê¤Ë¥Ð¥°¤¬¤¢¤ê, ¤¢¤ë¾ò·ï¤Ç¥æ¡¼¥¶¼½ñ¤È¥·¥¹¥Æ¥à¼½ñ¤¬Èó¸ß´¹¤Ë¤Ê¤ë¥Ð¥°¤ò½¤Àµ <li>¥æ¡¼¥¶¼½ñ¥Õ¥¡¥¤¥ë¤Îʸ»ú¥³¡¼¥É¤¬¥·¥¹¥Æ¥à¼½ñ¥Õ¥¡¥¤¥ë¤Îʸ»ú¥³¡¼ ¥É¤¬°Û¤Ê¤ë¾ì¹ç, ¼½ñ¤Î¹½ÃÛ¤¬¤¦¤Þ¤¯¤¤¤«¤Ê¤«¤Ã¤¿ÌäÂê¤Î½¤Àµ <li>¥³¥Þ¥ó¥É¥é¥¤¥ó¥ª¥×¥·¥ç¥ó¤ò¥À¥ó¥×¤¹¤ë --dump-config ¥ª¥×¥·¥ç¥ó¤ÎÄɲà <li>EM¥Ù¡¼¥¹¤ÎHMM³Ø½¬¤ò¥µ¥Ý¡¼¥È¤Ç¤¤ë¤è¤¦¤Ê³Ø½¬¥ë¡¼¥Á¥ó¤ÎÄɲà (experimental) </ul> <li><strong>2007-03-11</strong> MeCab 0.95<br> <ul> <li>¸Å¤¤¥³¥ó¥Ñ¥¤¥é¤Ç¥³¥ó¥Ñ¥¤¥ë¤Ç¤¤Ê¤¤ÌäÂê¤ò½¤Àµ <li>csv¤Î¥¨¥¹¥±¡¼¥×¤ÎÉÔ¶ñ¹ç¤Ç ","¤ò´Þ¤àñ¸ì¤¬ÄɲäǤ¤Ê¤«¤Ã¤¿ÌäÂê¤ò½¤Àµ <li>UTF8¼½ñ¤¬°ìÉôÀµ¾ï¤ËºîÀ®¤Ç¤¤Ê¤«¤Ã¤¿¥Ð¥°¤Î½¤Àµ <li>recall/precision¤Îɽ¼¨¤¬È¿ÂФˤʤäƤ¤¤¿¥Ð¥°¤Î½¤Àµ <li>¥³¥Þ¥ó¥É¥é¥¤¥ó²òÀϤÎÉÔ¶ñ¹ç¤Î½¤Àµ <li>¤½¤Î¾ºÙ¤«¤Ê¥Ð¥°¤Î½¤Àµ </ul> <li><strong>2007-02-24</strong>MeCab 0.94<br> <ul> <li>¿¤¯¤Î¥Ð¥°¥Õ¥£¥Ã¥¯¥¹ <li>HMM¤Ë¤è¤ë³Ø½¬¤ò¥µ¥Ý¡¼¥È (¼Â¸³Åª) <li>²òÀÏ·ë²Ì¤ÎÁ´¾ðÊó¤ò¼èÆÀ¤Ç¤¤ëAPI¤òÄɲà (begin_node_list, end_node_list) <li>char.def, unk.def, matrix.def ¤¬Ì¤ÄêµÁ¤Î¾ì¹ç¤Ç¤â¼½ñ¤¬ºîÀ®¤Ç¤¤ë¤è¤¦Êѹ¹ <li>WindowsÈǤΠiconv.dll¤Ø¤Î°Í¸¤òÇÑ»ß <li>¥³¡¼¥É¤Î¥¯¥ê¡¼¥ó¥¢¥Ã¥× </ul> <li><strong>2006-07-30</strong> MeCab 0.93<br> <ul> <li>¥é¥¤¥»¥ó¥¹¤òLGPL¤«¤éBSD,LGPL,GPL¤Î¥È¥ê¥×¥ë¥é¥¤¥»¥ó¥¹¤ËÊѹ¹ </ul> <li><strong>2006-07-10</strong> MeCab 0.92<br> <ul> <li>¼½ñ¥³¥ó¥Ñ¥¤¥éÅù, °ìÉôPerl¤Ç¼ÂÁõ¤µ¤ì¤Æ¤¤¤¿¥³¡¼¥É¤òC++¤ÇºÆ¼ÂÁõ. Perl¤Ø¤Î°Í¸À¤ÎÇÓ½ü <li>¼½ñ¥³¥ó¥Ñ¥¤¥é (mecab-dict-index) ¤Î¹â®²½ <li>rewrite.def ¤Î¥·¥ó¥¿¥Ã¥¯¥¹¤ÎÊѹ¹ <li>-x "̤ÃθìÉÊ»ì" ¥ª¥×¥·¥ç¥ó¤ÎÄɲÃ: ̤Ãθì¿äÄê¤ò¹Ô¤ï¤º, ¥æ¡¼¥¶¤¬»ØÄꤷ¤¿ "̤ÃθìÉÊ»ì" ¤ò½ÐÎÏ <li><a href="posid.html">ÉÊ»ì id</a> ¤Î¥µ¥Ý¡¼¥È <li>ʸ»ú¼ï¾ðÊ󤬰ìÉô³Ø½¬¤Ç¤¤Æ¤¤¤Ê¤«¤Ã¤¿¥Ð¥°¤Î½¤Àµ <li>³Ø½¬¤ÎºÝ, ÉÑÅ٤ˤè¤ëÂÀڤ꤬¤Ç¤¤Æ¤¤¤Ê¤«¤Ã¤¿¥Ð¥°¤Î½¤Àµ <li>¤½¤Î¾ºÙ¤¤¥Ð¥°¤Î½¤Àµ </ul> <li><strong>2006-04-30</strong> MeCab 0.91<br> <ul> <li>Windows ´Ä¶¤Çʸ»úÎó¤ÎºÇ¸å¤¬È¾³Ñ¥¹¥Ú¡¼¥¹¤Î»þ¤ËÍî¤Á¤ë¥Ð¥°¤Î½¤Àµ <li>Ï¢ÀÜɽ¤ÎÁ°·ï¤È¸å·ï¤Î¥µ¥¤¥º¤¬°Û¤Ê¤ë¤È¤¤ËÀµ¤·¤¯²òÀϤǤ¤Ê¤¤¥Ð¥° ¤Î½¤Àµ <li>mecab-dict-index ¤Ë -f ¥ª¥×¥·¥ç¥ó¤òÄɲä·, CSV ¤Îʸ»ú¥³¡¼¥É¤ò¥æ¡¼ ¥¶¤¬»ØÄê¤Ç¤¤ë¤è¤¦¤Ë¤·¤¿ <li>°ìÉô¤Î API´Ø¿ô¤¬ export ¤µ¤ì¤Æ¤¤¤Ê¤¤ÌäÂê¤Î½¤Àµ <li>CRF¤Î³Ø½¬¤ò pthread ¤ò»È¤Ã¤ÆÊÂÎó¤Ë¹Ô¤¨¤ë¤è¤¦¤Ë¤·¤¿ (experimental) <li>¥æ¡¼¥¶¼½ñ¤¬ºîÀ®¤Ç¤¤Ê¤¤ÌäÂê¤Î½¤Àµ <li>example ¥Ç¥£¥ì¥¯¥È¥ê¤Ë MeCab¤Î±þÍÑÎã¤òÄɲà (unittest) <li>¤½¤Î¾ºÙ¤¤¥Ð¥°¤Î½¤Àµ </ul> <li><strong>2006-03-26</strong> MeCab 0.90<br> <ul> <li>Initial release! </ul> </ul> <h2><a name="download">¥À¥¦¥ó¥í¡¼¥É</a></h2> <ul> <li><b>MeCab</b> ¤Ï¥Õ¥ê¡¼¥½¥Õ¥È¥¦¥§¥¢¤Ç¤¹¡¥<a href="http://www.gnu.org/licenses/gpl.txt">GPL</a>(the GNU General Public License), <a href="http://www.gnu.org/licenses/lgpl.txt">LGPL</a>(Lesser GNU General Public License), ¤Þ¤¿¤Ï BSD ¥é¥¤¥»¥ó¥¹¤Ë½¾¤Ã¤ÆËÜ¥½¥Õ¥È¥¦¥§¥¢¤ò»ÈÍÑ,ºÆÇÛÉÛ¤¹¤ë¤³¤È¤¬¤Ç¤¤Þ¤¹. ¾ÜºÙ¤Ï COPYING, GPL, LGPL, BSD³Æ¥Õ¥¡¥¤¥ë¤ò»²¾È¤·¤Æ²¼¤µ¤¤¡¥ <li><b>MeCab</b> ËÜÂÎ <h3><a name="source">Source</a></h3> <ul> <li>mecab-0.95.tar.gz:<a href="src">¥À¥¦¥ó¥í¡¼¥É</a> <li>¼½ñ¤Ï´Þ¤Þ¤ì¤Æ¤¤¤Þ¤»¤ó. Æ°ºî¤Ë¤ÏÊÌÅÓ¼½ñ¤¬É¬ÍפǤ¹. </ul> <h3><a name="win">Binary package for MS-Windows</a></h3> <ul> <li>mecab-0.95.exe:<a href="src">¥À¥¦¥ó¥í¡¼¥É</a> <li>Windows ÈÇ¤Ë¤Ï ¥³¥ó¥Ñ¥¤¥ëºÑ¤ß¤Î IPA ¼½ñ¤¬´Þ¤Þ¤ì¤Æ¤¤¤Þ¤¹</a> </li> </ul> <li><b>MeCab</b> ÍѤμ½ñ <h3>IPA ¼½ñ</h3> <ul> <li>IPA ¼½ñ, IPA¥³¡¼¥Ñ¥¹ ¤Ë´ð¤Å¤ <a href="http://www.cis.upenn.edu/~pereira/papers/crf.pdf">CRF</a> ¤Ç¥Ñ¥é¥á¡¼¥¿¿äÄꤷ¤¿¼½ñ¤Ç¤¹. <b>(¿ä¾©)</b> <a href="src">¥À¥¦¥ó¥í¡¼¥É</a></li> </ul> <h3>Juman ¼½ñ</h3> <ul> <li>Juamn ¼½ñ, µþÅÔ¥³¡¼¥Ñ¥¹¤Ë´ð¤Å¤ <a href="http://www.cis.upenn.edu/~pereira/papers/crf.pdf">CRF</a> ¤Ç¥Ñ¥é¥á¡¼¥¿¿äÄꤷ¤¿¼½ñ¤Ç¤¹. <a href="src">¥À¥¦¥ó¥í¡¼¥É</a></li> </ul> <h3>Canna dic</h3> <ul> <li>Canna ¼½ñ: ¸ø³«Í½Äê</li> </ul> <li><a name="script"><b>perl/ruby/python/java ¥Ð¥¤¥ó¥Ç¥£¥ó¥°</b></a> <ul> <li> <a href="src">¥À¥¦¥ó¥í¡¼¥É</a> </ul> </ul> </ul> <h2><a name="install">¥¤¥ó¥¹¥È¡¼¥ë</a></h2> <h3><a name="install-unix">UNIX</a></h3> <ul> <li>Æ°ºî¤ËɬÍפʤâ¤Î <ul> <li>C++ ¥³¥ó¥Ñ¥¤¥é (g++ 3.4.3 ¤È VC7 ¤Ç¤Î¥³¥ó¥Ñ¥¤¥ë¤ò³Îǧ¤·¤Æ¤¤¤Þ¤¹)</li> <li>iconv (libiconv): ¼½ñ¤Î¥³¡¼¥ÉÊÑ´¹¤Ë»È¤¤¤Þ¤¹</li> </ul> <li>¥¤¥ó¥¹¥È¡¼¥ë¼ê½ç <p>°ìÈÌŪ¤Ê¥Õ¥ê¡¼¥½¥Õ¥È¥¦¥§¥¢¤ÈƱ¤¸¼ê½ç¤Ç¥¤¥ó¥¹¥È¡¼¥ë¤Ç¤¤Þ¤¹.</p> <pre> % tar zxfv mecab-X.X.tar.gz % cd mecab-X.X % ./configure % make % make check % su # make install </pre> </li> <p>¼½ñ¤Î¥¤¥ó¥¹¥È¡¼¥ë</p> <pre> % tar zxfv mecab-ipadic-2.7.0-XXXX.tar.gz % mecab-ipadic-2.7.0-XXXX % ./configure % make % su # make install </pre> </ul> </ul> <h3><a name="install-windows">Windows</a></h3> <p>¥Ð¥¤¥Ê¥ê¤ò¥¤¥ó¥¹¥È¡¼¥ë¤¹¤ë¾ì¹ç¤Ï, ¼«¸Ê²òÅ।¥ó¥¹¥È¡¼¥é (mecab-X.X.exe) ¤ò¼Â¹Ô¤·¤Æ¤¯¤À¤µ¤¤. ¼½ñ¤âƱ»þ¤Ë¥¤¥ó¥¹¥È¡¼¥ë¤µ¤ì¤Þ¤¹.</p> <h2><a name="usage-tools">»È¤¤Êý</a></h2> <h3><a name="parse">¤È¤ê¤¢¤¨¤º²òÀϤ·¤Æ¤ß¤ë</a></h3> <p>mecab ¤òµ¯Æ°¤·¤Æ, À¸Ê¸¤òɸ½àÆþÎϤ«¤éÆþÎϤ·¤Æ¤ß¤Æ¤¯¤À¤µ¤¤.MeCab ¤Ç¤Ï, °ì¹Ô°ìʸ¤òÁ°Äó¤È¤·¤Æ²òÀϤò¹Ô¤Ê¤¤¤Þ¤¹.</p> <pre> % mecab ¤¹¤â¤â¤â¤â¤â¤â¤â¤â¤Î¤¦¤Á ¤¹¤â¤â ̾»ì,°ìÈÌ,*,*,*,*,¤¹¤â¤â,¥¹¥â¥â,¥¹¥â¥â ¤â ½õ»ì,·¸½õ»ì,*,*,*,*,¤â,¥â,¥â ¤â¤â ̾»ì,°ìÈÌ,*,*,*,*,¤â¤â,¥â¥â,¥â¥â ¤â ½õ»ì,·¸½õ»ì,*,*,*,*,¤â,¥â,¥â ¤â¤â ̾»ì,°ìÈÌ,*,*,*,*,¤â¤â,¥â¥â,¥â¥â ¤Î ½õ»ì,Ï¢Âβ½,*,*,*,*,¤Î,¥Î,¥Î ¤¦¤Á ̾»ì,Èó¼«Î©,Éû»ì²Äǽ,*,*,*,¤¦¤Á,¥¦¥Á,¥¦¥Á EOS </pre> <p> ½ÐÎÏ¥Õ¥©¡¼¥Þ¥Ã¥È¤Ï, ChaSen ¤Î¤½¤ì¤ÈÂ礤¯°Û¤Ê¤ê¤Þ¤¹. º¸¤«¤é, </p> <pre> ɽÁØ·Á\tÉÊ»ì,ÉÊ»ìºÙʬÎà1,ÉÊ»ìºÙʬÎà2,ÉÊ»ìºÙʬÎà3,³èÍÑ·Á,³èÍÑ·¿,¸¶·Á,Æɤß,ȯ²» </pre> <p>¤È¤Ê¤Ã¤Æ¤¤¤Þ¤¹. </p> <p>°ú¿ô¤Ë¥Õ¥¡¥¤¥ë¤òÍ¿¤¨¤ë¤È, ¤½¤Î¥Õ¥¡¥¤¥ë¤¬²òÀÏÂоݤȤʤê¤Þ¤¹. ¤Þ¤¿, -o ¥ª¥×¥·¥ç¥ó¤Ë¤Æ, Ê̤Υե¡¥¤¥ë¤Ë·ë²Ì¤ò½ÐÎϤ¹¤ë¤³¤È¤â²Äǽ¤Ç¤¹.</p> <pre> % mecab INPUT -o OUTPUT </pre> <h3><a name="wakati">¤ï¤«¤Á½ñ¤¤ò¤¹¤ë</a></h3> <p>°Ê²¼¤Î¤è¤¦¤Ë -O ¥ª¥×¥·¥ç¥ó¤ò»È¤¤¤Þ¤¹.</p> <pre> % mecab -O wakati ÂÀϺ¤Ï¤³¤ÎËܤòÆóϺ¤ò¸«¤¿½÷À¤ËÅϤ·¤¿¡£ ÂÀϺ ¤Ï ¤³¤Î ËÜ ¤ò ÆóϺ ¤ò ¸« ¤¿ ½÷À ¤Ë ÅϤ· ¤¿ ¡£ </pre> <h3><a name="format">½ÐÎÏ¥Õ¥©¡¼¥Þ¥Ã¥È¤ÎÊѹ¹</a></h3> <p>°Ê²¼¤Î¤è¤¦¤Ë -O ¥ª¥×¥·¥ç¥ó¤ò»È¤¤¤Þ¤¹.</p> <pre> % mecab -Oyomi (¥è¥ßÉÕÍ¿) % mecab -Ochasen (ChaSen¸ß´¹) % mecab -Odump (Á´¾ðÊó¤ò½ÐÎÏ) </pre> <p> ¤³¤ì¤é¤Î½ÐÎÏ¥Õ¥©¡¼¥Þ¥Ã¥È¤Ï, /usr/local/lib/mecab/ipadic/dicrc ¤ËÄêµÁ¤µ¤ì¤Æ¤¤¤Þ¤¹. ¤µ¤é¤Ë, ¥æ¡¼¥¶¤¬¤³¤ì¤é¤Î¥Õ¥©¡¼¥Þ¥Ã¥È¤ò¼«Í³¤ËÄêµÁ¤¹¤ë¤³¤È¤¬²Äǽ¤Ç¤¹. <a href="format.html">¤³¤Á¤é</a>¤ò¤´Í÷¤¯¤À¤µ¤¤.</p> <h2><a name="usage-tools2">¹âÅ٤ʻȤ¤Êý</a></h2> <h3><a name="charset">ʸ»ú¥³¡¼¥ÉÊѹ¹</a></h3> <p>Æä˻ØÄꤷ¤Ê¤¤¸Â¤ê, euc ¤¬»ÈÍѤµ¤ì¤Þ¤¹. ¤â¤·, shift-jis ¤ä utf8 ¤ò »È¤¤¤¿¤¤¾ì¹ç¤Ï, ¼½ñ¤Î configure ¥ª¥×¥·¥ç¥ó¤Ë¤Æ charset ¤òÊѹ¹¤·, ¼½ñ¤òºÆ¹½ÃÛ¤·¤Æ¤¯¤À¤µ¤¤. ¤³¤ì¤Ç, shift-jis ¤ä, utf8 ¤Î¼½ñ¤¬ºîÀ®¤µ¤ì¤Þ¤¹.</p> <pre> % tar zxfv mecab-ipadic-2.7.0-xxxx % cd mecab-ipadic-2.7.0-xxxx % ./configure --with-charset=sjis % make % tar zxfv mecab-ipadic-2.7.0-xxxx % ./configure --with-charset=utf8 % make </pre> <p>¤Þ¤¿, mecab-dict-index ¤Î -t ¥ª¥×¥·¥ç¥ó¤ò»È¤Ã¤ÆľÀÜʸ»ú¥³¡¼¥É¤Î°Û¤Ê¤ë ¼½ñ¤òºÆ¹½ÃۤǤ¤Þ¤¹. -f ¥ª¥×¥·¥ç¥ó¤Ï¥ª¥ê¥¸¥Ê¥ë¤Î¥Æ¥¥¹¥È¼½ñ¤Îʸ»ú¥³¡¼¥É¤Ç¤¹. </p> <pre> % cd mecab-ipadic-2.7.0-xxxx % /usr/local/libexec/mecab/mecab-dict-index -f euc-jp -t utf-8 # make install </pre> </pre> <h3><a name="utf-8">UTF-8 only mode</a></h3> <p> configure option ¤Ç --enable-utf8-only ¤ò»ØÄꤹ¤ë¤È. MeCab ¤¬°·¤¦ ʸ»ú¥³¡¼¥É¤ò utf8 ¤Ë¸ÇÄꤷ¤Þ¤¹. euc-jp ¤ä shift-jis ¤ò¥µ¥Ý¡¼¥È¤¹¤ë¾ì¹ç, MeCab ÆâÉô¤ËÊÑ´¹ÍѤΥơ¼¥Ö¥ë¤òËä¤á¤³¤ß¤Þ¤¹. --enable-utf8-only ¤ò »ØÄꤹ¤ë¤³¤È¤Ç¥Æ¡¼¥Ö¥ë¤ÎËä¤á¤³¤ß¤òÍÞÀ©¤·, ·ë²Ì¤È¤·¤Æ¼Â¹Ô¥Ð¥¤¥Ê¥ê¤ò ¾®¤µ¤¯¤¹¤ë¤³¤È¤¬¤Ç¤¤Þ¤¹.</p> <h3><a name="unk">̤Ãθì¿äÄê</a></h3> <p> MeCab ¤Ï, ¼½ñ¤Ëñ¸ì¤¬Ì¤ÅÐÏ¿¤Î¾ì¹ç¤Ç¤âŬÅö¤Ë¤½¤ÎÉÊ»ì¤ò¿äÄꤷ¤Þ¤¹. </p> <pre> ¥Û¥ê¥¨¥â¥ó»Ô ¥Û¥ê¥¨¥â¥ó ̾»ì,¸ÇÍ̾»ì,ÃÏ°è,°ìÈÌ,*,*,* »Ô ̾»ì,ÀÜÈø,ÃÏ°è,*,*,*,»Ô,¥·,¥· EOS ¥Û¥ê¥¨¥â¥ó¤µ¤ó ¥Û¥ê¥¨¥â¥ó ̾»ì,¸ÇÍ̾»ì,¿Í̾,°ìÈÌ,*,*,* ¤µ¤ó ̾»ì,ÀÜÈø,¿Í̾,*,*,*,¤µ¤ó,¥µ¥ó,¥µ¥ó </pre> <p>¤¿¤À¤·, ¤½¤ÎÀºÅÙ¤ÏÀµ³Î¤Ç¤Ï¤¢¤ê¤Þ¤»¤ó. ÉÊ»ì¿äÄê¤ò¤ä¤á, ̤Ãθì¤Ï¾ï¤Ë "̤Ãθì" ÉÊ»ì¤ò½ÐÎϤ·¤¿¤¤¾ì¹ç¤Ï -x (--unk-feature) ¥ª¥×¥·¥ç¥ó¤ò»È¤¤¤Þ¤¹. ¥ª¥×¥·¥ç¥ó¤Ç»ØÄꤵ¤ì¤¿Ê¸»úÎó¤¬ÉÊ»ì¤È¤·¤Æ»È¤ï¤ì¤Þ¤¹.</p> </p> <pre>%mecab --unk-feature "̤Ãθì" ¥Û¥ê¥¨¥â¥ó¤µ¤ó ¥Û¥ê¥¨¥â¥ó ̤ÃÎ¸ì ¤µ¤ó ̾»ì,ÀÜÈø,¿Í̾,*,*,*,¤µ¤ó,¥µ¥ó,¥µ¥ó </pre> </p> <h3><a name="nbest">N-Best ²ò¤Î½ÐÎÏ</a></h3> <p> -N #NUM ¥ª¥×¥·¥ç¥ó¤ò»È¤¦¤³¤È¤Ç, ³Î¤«¤é¤·¤¤¤â¤Î¤«¤é#NUM ¸Ä²òÀÏ·ë²Ì¤ò½ÐÎÏ ¤·¤Þ¤¹. ÍýÏÀŪ¤Ë¤Ï¤¹¤Ù¤Æ¤Î²Äǽ¤Ê²òÀϲò¤ò½ÐÎϤ¹¤ë¤³¤È¤¬ ²Äǽ¤Ç¤¹¤¬, ½ÐÎϥХåե¡¤Î¤«¤Í¤¢¤¤¤«¤é, -N ¤ÎºÇÂçÃͤò 512 ¤ËÀ©¸Â¤·¤Æ¤¤¤Þ¤¹. </p> <pre> % mecab -N2 º£Æü¤â¤·¤Ê¤¤¤È¤Í¡£ º£Æü ̾»ì,Éû»ì²Äǽ,*,*,*,*,º£Æü,¥¥ç¥¦,¥¥ç¡¼ ¤â ½õ»ì,·¸½õ»ì,*,*,*,*,¤â,¥â,¥â ¤· Æ°»ì,¼«Î©,*,*,¥µÊÑ¡¦¥¹¥ë,̤Á³·Á,¤¹¤ë,¥·,¥· ¤Ê¤¤ ½õÆ°»ì,*,*,*,Æü졦¥Ê¥¤,´ðËÜ·Á,¤Ê¤¤,¥Ê¥¤,¥Ê¥¤ ¤È ½õ»ì,Àܳ½õ»ì,*,*,*,*,¤È,¥È,¥È ¤Í ½õ»ì,½ª½õ»ì,*,*,*,*,¤Í,¥Í,¥Í ¡£ µ¹æ,¶çÅÀ,*,*,*,*,¡£,¡£,¡£ EOS º£Æü ̾»ì,Éû»ì²Äǽ,*,*,*,*,º£Æü,¥¥ç¥¦,¥¥ç¡¼ ¤â¤· Éû»ì,°ìÈÌ,*,*,*,*,¤â¤·,¥â¥·,¥â¥· ¤Ê¤¤ ·ÁÍÆ»ì,¼«Î©,*,*,·ÁÍƻ졦¥¢¥¦¥ªÃÊ,´ðËÜ·Á,¤Ê¤¤,¥Ê¥¤,¥Ê¥¤ ¤È ½õ»ì,Àܳ½õ»ì,*,*,*,*,¤È,¥È,¥È ¤Í ½õ»ì,½ª½õ»ì,*,*,*,*,¤Í,¥Í,¥Í ¡£ µ¹æ,¶çÅÀ,*,*,*,*,¡£,¡£,¡£ EOS </pre> <h2><a name="thanks">¼Õ¼</a></h2> <p> CRF ¤Î¥Ñ¥é¥á¡¼¥¿¿äÄê¤Ë <a href="http://www.ece.nwu.edu/~nocedal">Jorge Nocedal</a> »á¤¬¹Í°Æ¤·¤¿ L-BFGS ¤ÈƱ»á¤¬¸ø³«¤·¤Æ¤¤¤ë FORTRAN ¼ÂÁõ¤ò»È¤ï¤»¤Æ¤¤¤¿¤À¤¤¤Æ¤ª¤ê¤Þ¤¹¡£¤¢¤ê¤¬¤È¤¦¤´¤¶¤¤¤Þ¤¹.</p> <p><a href="http://www.ece.northwestern.edu/~nocedal/lbfgs.html">http://www.ece.northwestern.edu/~nocedal/lbfgs.html</a></p> <ul> <li> J. Nocedal. Updating Quasi-Newton Matrices with Limited Storage (1980), Mathematics of Computation 35, pp. 773-782. <li>D.C. Liu and J. Nocedal. On the Limited Memory Method for Large Scale Optimization (1989), Mathematical Programming B, 45, 3, pp. 503-528. </ul> </p> <hr> <p>$Id: index.html 135 2007-06-10 12:28:13Z taku-ku $;</p> </body> </html>