A Basque stemming algorithm


 

Links to resources

Snowball main page
Tar-gzipped file of the delivered resources
Basque suffixes (as referenced below)


A stemmer for the Basque (Euskera) language was sent us by Mikel Otxandorena of iSOCO with this accompanying email:
From: Mikel Otxandorena Date: Mon, 26 Jul 2010 13:04:19 +0200 Subject: Basque Snowball Stemmer Hi, As I told you the other day, I am sending you a valid version of a Basque stemmer. I have tested it with a dictionary that I attach in the email and it works well. For our use of suffixes see the following table: http://www.santurtzieus.com/gelairekia/unitateak/3/lagun/atzizkien_zerrenda.htm We are open to feedback and discussion about the linguistics, and about the final form of the snowball implementation. Please distribute this freely, with a mention of the author and of the company iSOCO, that has let us share our work freely with the Snowball community. Kind Regards, Mikel Otxandorena
Mikel has confirmed that he wishes iSOCO to be duly acknowledged on the Snowball site, and that the stemmer itself is free for use under the BSD license.

Here is the script of the stemmer in Snowball,


routines ( aditzak izenak adjetiboak mark_regions RV R2 R1 ) externals ( stem ) integers ( pV p1 p2 ) groupings ( v ) stringescapes {} /* special characters (in ISO Latin I) */ stringdef n~ hex 'f1' define v 'aeiou' define mark_regions as ( $pV = limit $p1 = limit $p2 = limit // defaults do ( ( v (non-v gopast v) or (v gopast non-v) ) or ( non-v (non-v gopast v) or (v next) ) setmark pV ) do ( gopast v gopast non-v setmark p1 gopast v gopast non-v setmark p2 ) ) backwardmode ( define RV as $pV <= cursor define R2 as $p2 <= cursor define R1 as $p1 <= cursor define aditzak as ( [substring] among( 'le' 'la' 'tzaile' 'aldatu' 'atu' 'tzailea' 'taile' 'tailea' 'pera' 'gale' 'galea' 'gura' 'kura' 'kor' 'korra' 'or' 'orra' 'tun' 'tuna' 'gaitz' 'gaitza' 'kaitz' 'kaitza' 'ezin' 'ezina' 'tezin' 'tezina' 'errez' 'erreza' 'karri' 'karria' 'tzaga' 'tzaka' 'tzake' 'tzeke' 'ez' 'eza' 'tzez' 'keta' 'eta' 'etan' 'pen' 'pena' 'tze' 'atze' 'kuntza' 'kunde' 'kundea' 'kune' 'kunea' 'kuna' 'kera' 'era' 'kizun' 'kizuna' 'dura' 'tura' 'men' 'mena' 'go' 'ago' 'tio' 'taldi' 'taldia' 'aldi' 'aldia' 'gune' 'gunea' 'bide' 'bidea' 'pide' 'pidea' 'gai' 'gaia' 'ki' 'kin' 'rekin' 'kina' 'kari' 'karia' 'ari' 'tari' 'etari' 'gailu' 'gailua' 'kide' 'kidea' 'ide' 'idea' 'du' 'ka' 'kan' 'an' 'ean' 'tu' 'lari' 'tatu' 'rean' 'tarazi' 'arazi' 'tzat' 'bera' 'dako' ( RV delete ) 'garri' 'garria' 'tza' (R2 delete) 'atseden' (<- 'atseden') 'arabera' (<- 'arabera') 'baditu' (<- 'baditu') ) ) define izenak as ( [substring] among( 'ari' 'aria' 'bizia' 'kari' 'karia' 'lari' 'laria' 'tari' 'taria' 'zain' 'zaina' 'tzain' 'tzaina' 'zale' 'zalea' 'tzale' 'tzalea' 'aizun' 'orde' 'ordea' 'burua' 'ohi' 'ohia' 'kintza' 'gintzo' 'gintzu' 'tzu' 'tzua' 'tzo' 'tzoa' 'kuntza' 'talde' 'taldea' 'eria' 'keria' 'teria' 'di' 'za' 'ada' 'tara' 'etara' 'tra' 'ta' 'tegi' 'tegia' 'keta' 'z' 'zko' 'zkoa' 'ti' 'tia' 'tsu' 'tsua' 'zu' 'zua' 'bera' 'pera' 'zto' 'ztoa' 'asi' 'asia' 'gile' 'gilea' 'estu' 'estua' 'larri' 'larria' 'nahi' 'nahia' 'koi' 'koia' 'oi' 'oia' 'goi' 'min' 'mina' 'dun' 'duna' 'duru' 'durua' 'duri' 'duria' 'os' 'osa' 'oso' 'osoa' 'ar' 'ara' 'tar' 'dar' 'dara' 'tiar' 'tiara' 'liar' 'liara' 'gabe' 'gabea' 'kabe' 'kabea' 'ga' 'ge' 'kada' 'tasun' 'tasuna' 'asun' 'asuna' 'go' 'mendu' 'mendua' 'mentu' 'mentua' 'mendi' 'mendia' 'zio' 'zioa' 'zino' 'zinoa' 'zione' 'zionea' 'ezia' 'degi' 'degia' 'egi' 'egia' 'toki' 'tokia' 'leku' 'lekua' 'gintza' 'alde' 'aldea' 'kalde' 'kaldea' 'gune' 'gunea' 'une' 'unea' 'una' 'pe' 'pea' 'gibel' 'gibela' 'ondo' 'ondoa' 'arte' 'artea' 'aurre' 'aurrea' 'etxe' 'etxea' 'ola' 'ontzi' 'ontzia' 'gela' 'denda' 'taldi' 'taldia' 'aldi' 'aldia' 'te' 'tea' 'zaro' 'zaroa' 'taro' 'taroa' 'oro' 'oroa' 'aro' 'aroa' 'ero' 'eroa' 'eroz' 'eroza' 'ka' 'kan' 'kana' 'tako' 'etako' 'takoa' 'kote' 'kotea' 'tzar' 'tzarra' 'handi' 'handia' 'kondo' 'kondoa' 'skila' 'no' 'noa' '{n~}o' '{n~}oa' 'ska' 'xka' 'zka' 'tila' 'to' 'toa' 'tto' 'ttoa' 'txo' 'txoa' 'txu' 'txua' 'anda' 'anga' 'urren' 'urrena' 'gai' 'gaia' 'gei' 'geia' 'eme' 'emea' 'kume' 'kumea' 'sa' 'ko' 'eko' 'koa' 'ena' 'enea' 'ne' 'nea' 'kor' 'korra' 'ez' 'eza' 'eta' 'etan' 'ki' 'kia' 'kin' 'kina' 'tu' 'tua' 'du' 'dua' 'ek' 'tarik' 'tariko' 'tan' 'ordu' 'ordua' 'oste' 'ostea' 'tzara' 'ra' 'antza' 'behar' 'ro' 'giro' 'ak' 'zp' 'ket' 'kail' 'kaila' 'ail' 'kirri' 'kirria' 'ngo' 'ngoa' '{n~}i' 'sko' 'sta' 'koitz' 'koitza' 'na' 'garren' 'garrena' 'kera' 'gerren' 'gerrena' 'garna' 'kide' 'tz' 'tuko' ( RV delete ) 'ora' 'garri' 'garria' 'or' 'buru' 'ren' 'tza' ( R2 delete ) 'joka' (<- 'jok') 'tzen' 'ten' 'en' 'tatu' (R1 delete) 'trako' (<- 'tra') 'minutuko' (<- 'minutu') 'zehar' (<- 'zehar') 'geldi' (<- 'geldi') 'igaro' (<- 'igaro') 'aurka' (<- 'aurka') ) ) define adjetiboak as ( [substring] among( 'era' 'ero' 'go' 'tate' 'tade' 'date' 'dade' 'keria' 'ki' 'to' 'ro' 'la' 'gi' 'larik' 'lanik' 'ik' 'ztik' 'rik' ( RV delete ) 'zlea' (<- 'z') ) ) ) define stem as ( do mark_regions backwards ( repeat aditzak repeat izenak do adjetiboak ) ) /* Note 1: additions of 21 Jul 2010 */