|
|
|
|
|
|
|
|
home
|
about
|
usage
|
demo
|
|
|
|
|
Dictionary |
80000 vocabulary entries of the Swedish dictionary
(based on the "Nationalencyklopedins ordbok").
|
|
|
Basic glossaries of English, French, German, Latin.
|
|
|
50000 proper names: people, places, organisations, etc.
|
|
|
Representation of Swedish inflection corresponding to 800000 forms.
|
|
|
75 inflection descriptions
11 parts of speech
48 form types.
|
|
|
Links among lexical units: hypo/hypernymy, synonymy, derviations, compound components.
|
|
|
Specification of: ethymology, usage and topic
(about 100 topic categories).
|
|
|
Grammar includes 400 word formation rules, 500 general phrase rules and 700 collocational patterns.
|
|
|
Linguistic engine
|
Lexware development begun in 1995, it is a registered trade mark since 2000.
|
|
|
Already in the beginning the engine is quick - about 20000 words are analyzed in a second on a PIII 800Mhz
(lazy analysis strategy).
|
|
|
98% coverage is obtained with unique recognition. |
|
|
Analysis engine is written in standard C, applications in standard Java, hence the system is fully portable. |
|
|
Best in external evaluation of a text categorization application for documents of the Swedish Parliament - Djupindexering.
|
|
|
Djupindexering
|
A text categorization application which exploits Lexware's possiblities to integrate its knowledge of Swedish with a thesaurus from the Library of the Swedish Parliament.
|