Lodestar.Text 0.6.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

DanishSnowballStemmer

Danish stemming by the Snowball algorithm.

public static class DanishSnowballStemmer

Example — a noun, its definite form and the verb it comes from, collapsed onto one key.

using Lodestar.Text.Stemming;

string noun = DanishSnowballStemmer.Stem("bestemmelse");  // => bestem
string definite = DanishSnowballStemmer.Stem("bestemmelsen");  // => bestem
string verb = DanishSnowballStemmer.Stem("bestemmer");  // => bestem

Remarks — Danish reads as Swedish with one step added: four steps rather than three, all of them in R1, with no R2 and no RV region. Three things about it are worth knowing before reading a stem.

The fourth step undoubles. A word left ending in bb, dd, ff, gg, kk, ll, mm, nn, pp, rr, ss or tt loses the second letter, and it runs last — after the deletions that uncover the pair. bestemmelse loses its e, then its els, reaches bestemm, and only then becomes bestem. This is the step Swedish does not have, and the reason a Danish stem is often one letter shorter than it looks like it should be.

The third step runs the second one again. Deleting ig, lig, elig or els can uncover a gd, dt, gt or kt that was not there before, so the consonant-pair step is repeated afterwards. fjendtligt loses its t to step 2, its lig to step 3, and then its second t to step 2 a second time, ending at fjend.

The region is part of the search, not a test after it. As in Swedish, each step looks for the longest suffix that lies in R1 — not the longest suffix, checked against R1 afterwards. nyheder ends with heder, which reaches past R1, and with er, which does not; er is the one that goes, and the word stems to nyhed. frihed comes back whole for the same reason.

R1 is floored at three characters as in Swedish and German, which is what stops short words from being trimmed to nothing. virus also comes back whole, but for the other reason: its s follows a u, which is not one of the twenty letters the algorithm accepts before a bare s.

The published description gives R1 a second definition for words containing an apostrophe, which nltk does not implement and neither does this — see decision 0087.

Reference behaviour is nltk.stem.snowball.SnowballStemmer("danish"), matched over 240 words.

Applies to — net10.0, netstandard2.0.

See also — the stemming index.

Members

Member	What it does
DanishSnowballStemmer.Stem	The Snowball stem of one Danish word.