Project: syntactically analyzed and disambiguated text corpus

Project "Syntactically analyzed and disambiguated text corpus"

financed by Ministry of Education and Research of Estonia program "Estonian language and national culture"
Duration 2002 - 2003

The aim of the project is to extend the testing corpus for the Constraint Grammar shallow syntactic parser, written by K. Müürisep and K. Puolakainen, to gain better precision and recall of the parser. The corpus consists of fiction (107 000 words¹) + newspaper texts (10 000 words) + legal texts (6 000 words). In 2003, we are about to annotate 70 000 words of fiction. By the end of the year 2003, the size of the test corpus will be 200 000 words.

Project leader: Tiit Roosmaa, PhD

Staff:

"senior"

Kaili Müürisep, PhD (computer science)
Kadri Muischnek, M.A. (general linguistics)
Heli Uibo, M.Sc. (computer science)
Heili Orav, M.A. (general linguistics)
Andriela Rääbis, M.A. (Estonian language)

"junior" (students of computational linguistics):

Siiri Pärkson
Helen Nigol
Kadri Kerner
Birge Talve

Results:

Report 2002 (in Estonian)
Directory of shallow syntactically annotated texts
Syntactic Relations' Dictionary by K. Kaljurand (work in progress), derived from the corpus

Useful links for syntactic annotators:

Instructions for hand-annotators by K. Müürisep (in Estonian)
Hand-annotator's tool by K. Pirn (in Estonian)

¹ - state of January 2003
The page is maintained by Heli Uibo (heli_u@ut.ee). Last modified Feb 27, 2004.