specializes in natural language processing (NLP). Main product is the linguistic engine EXTRAKT which serves as an add-on to any kind of search system. Most of the European languages including
German are covered.
With the help of EXTRAKT monolingual or multilingual (cross-lingual) search systems can be created.
EXTRAKT is an integrated system of linguistic and statistical functions for
- the identification of basic forms (Lemmatisation),
- the splitting of compound words (German and Dutch),
- the generation of inflexed forms (Generation),
- the access to synonyms, derivated words and thesauri,
- the translation of terms into other languages,
- the phonetic transformation of (proper) names or plain words for a fuzzy search (TRAPHO),
- the generation of syntactical correct word groups (phrasal generation),
- the identification of all variants of a stem of a given word (linguistic stemming),
- identification of sentence borders, word groups and (Named Entity Extraction ).
EXTRAKT in fact is a complete linguistic platform and is working in a multitude of different applications of our partners.
Complementary to the linguistic engine EXTRAKT is the BESTWORD server for a fault tolerant search.
BESTWORD is able to identify via a combined phonetic/graphemic algorithm to find similar relevant alternatives to a given word - out of millions of entries in a couple of milli-seconds.