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

FinnishSnowballStemmer.Stem

The Snowball stem of one Finnish word.

public static string Stem(string word)

Parameters — word is a single Finnish word. It is lowercased and NFC-normalised before the rules run, so a decomposed ä and a composed one are the same letter to the algorithm.

Returns — string, the stem, always lowercase, and still carrying whatever ä or ö the input had.

Exceptions — ArgumentNullException when word is null. An empty string, or a word of one character, is returned lowercased and otherwise untouched.

Example — a plural, a partitive and a compound illative.

using Lodestar.Text.Stemming;

string plural = FinnishSnowballStemmer.Stem("talot");  // => talo
string partitive = FinnishSnowballStemmer.Stem("taloja");  // => talo
string illative = FinnishSnowballStemmer.Stem("kotimaahan");  // => kotim

Remarks — kotimaahan → kotim is the result most likely to look like a bug. It is not. The illative -han goes with the third step, which leaves kotimaa; the tidying step then removes the long vowel aa, leaving kotima, and then the consonant-plus-a at the end, leaving kotim. Each of those runs at most once, in turn. A stemmer produces keys, not words.

The same step explains the other short answers. kissaa loses its partitive -a, then its sa, then one of its doubled s, and ends at kis.

The -sti adverb is the one ending in the first step that has to sit in R2 rather than R1, and that is the whole difference between the two results below: nopeasti is too short for it, so only the tidying step touches the word.

using Lodestar.Text.Stemming;

string shortAdverb = FinnishSnowballStemmer.Stem("nopeasti");  // => nopeast
string longAdverb = FinnishSnowballStemmer.Stem("todellisesti");  // => todellis
string genitive = FinnishSnowballStemmer.Stem("huoneen");  // => huone

Applies to — net10.0, netstandard2.0.

See also — FinnishSnowballStemmer, the stemming index.