1 st Workshop on Intelligent and Knowledge oriented Technologies

Levenshtein edit operation is a basic string operation – insertion, deletion or substitution of a character in a string. Sequence of edit operations can be used to transform basic word form (lemma) into an inflected form, and the same sequence can can be used to transform lemmata belonging to the same inflectional paradigm. Presented system contains inflection paradigms of over 56000 lemmata from Short Dictionary of Slovak Language and from the most frequent word forms in the Slovak National Corpus, together with detailed grammar information about each generated word form. 1 Levenshtein distance and some definitions Levenshtein distance[1] is a metric defined on the space of strings as a minimum number of Levenshtein edit operations needed to transform one string into the other, where by a Levenshtein edit operation we understand insertion, deletion or a substitution of a character. A Levenshtein edit operation e can be formally described as e = (o, s, d) – a triple of operation type o, position in the source string s and position in the destination string d, where operation type o is one of replace, insert or delete. For replace or insert, the replacement/new character is taken from the destination string. Sequence of edit operations q = (e1, e2, e3, ...), together with the destination string D, when applied to a string S ∈ S defines a mapping function f : S 7→ S, where S is a set of all strings. To each word form w ∈ W, where W is a set of all the words we can assign a set of grammar categories Gw = {g1, g2, g3, ...} represented by short mnemotechnical strings (called morphological tags). Now for each tagged word form together with its morphological tag (wi, gi) ∈ W × G there exists a mapping function fi consisting of Levenshtein edit operations such that fi(l) = wi, where l is a deliberately chosen word, called lemma and considered to be a basic word form for a given lexeme.