A prototype morphological-semantical parser for English compound participles
is now available at http://jens.zorba.dk/phd/cpparser_2.htm
Any comments, criticisms, suggestions for improvements, etc. are most
welcome, esp. from native English speakers.
Introduction:
CP-Parser 2 (CP2) parses English NP constructions of the following syntax:
NP ::= <prefix> + ‘-‘ + <participle> + ‘ ‘ + <head>
Prefix ::= <noun> | <adj> | <adv> | <prep> | <LBM>
Participle ::= <present participle> | <past participle>
Head ::= <noun>
where LBM = lexical, bound morpheme.
CP2 includes approx. 100 present and 100 past participle examples, which
were extracted from the British National Corpus (BNC), using the Corpus
Query Processor (CQP) tool © IMS, Stuttgart University.
The BNC data provides the core of CP2’s lexicon, which was formatted by
means of the WordSmith application. The word lists were then tagged and
lemmatized by Conexor’s web tagger at www.conexor.fi. Finally, information
concerning valency, semantic selectional restrictions and semantic
categorization was added manually.
The morphological-semantical parsing algorithm builds on the principles set
out at http://jens.zorba.dk/phd - further documentation is forthcoming.
Mange Hilsener
Jens Ahlmann Hansen
Karlsbjergvej 39 C
DK-5672 Broby
mailto:jahlmann@mail.tele.dk
TEL: 00 45 65 50 33 10 (office)
00 45 62 63 39 39 (home)
00 45 29 46 43 22 (mobile)
This archive was generated by hypermail 2b29 : Wed Mar 21 2001 - 13:36:55 MET