Shallow parsing of Hungarian business news
Tamás Váradi · 2003
The present paper reports on an attempt to annotate noun phrases in Hungarian using cascaded regular grammars. Hungarian presents several difficulties to shallow parsing such as discourse oriented constituent order as well as left-branching recursive possessive and participle structure inside noun phrases. The approach uses cascaded regular grammars and was developed with the CLaRK system. The NP grammar was tested on a morphologically tagged and disambiguated corpus of 928 sentences representing a sample of highly sophisticated written style of journalism. The results are encouraging even on this challenging text type.