Development build. This page describes main, not a released package. The latest published Lodestar.Text is 0.6.0 — read its documentation.

Home › Text › Stemming

DutchSnowballStemmer

Dutch stemming by the Snowball algorithm.

public static class DutchSnowballStemmer

Example — a noun and its plural, collapsed onto one key.

using Lodestar.Text.Stemming;

string singular = DutchSnowballStemmer.Stem("boek");  // => boek
string plural = DutchSnowballStemmer.Stem("boeken");  // => boek
string diminutive = DutchSnowballStemmer.Stem("boekje");  // => boekj

Remarks — Dutch is built like the German stemmer next door and differs from it in three places worth knowing before reading a stem.

Accents are folded before the rules run, not after. ä ë ï ö ü and á é í ó ú become their bare vowel first, so één stems as een does. The grave on è is the exception: it is a letter of the Dutch alphabet rather than a diacritic, counts as a vowel, and survives.

A stressed vowel is undoubled at the end. maan and manen are one word spelled two ways, and the last step is what makes them one key: a doubled aa, ee, oo or uu between two consonants loses a letter, so both come back as man.

And there is no RV region — R1 and R2 only, with R1 floored at three characters as in German. That floor is what stops short words from being trimmed to nothing, the job RV does in the Romance languages.

Derivational endings go further than inflectional ones. wandeling, wandelingen and wandelend all reach wandel, but only because ing and end sit inside R2; lopend is too short for that and comes back whole.

Reference behaviour is nltk.stem.snowball.SnowballStemmer("dutch"), matched over 177 words.

Applies to — net10.0, netstandard2.0.

See also — the stemming index.

Members

Member	What it does
DutchSnowballStemmer.Stem	The Snowball stem of one Dutch word.