Automatically constructed multiword lexicon hrMWELex v0.5

Nikola Ljubešić · Americanae (AECID Library) · 2015

The hrMWELex lexicon is an automatically constructed lexicon of Croatian multiword expression candidates (mostly collocations) from the parsed hrWaC 2.0 corpus by using the DepMWEx [depmueks] tool (https://github.com/nljubesi/depmwex). The tool extracts MWE candidates from parse trees by applying tree patterns and ranking by occurrence statistics.

Read the paper · More papers on PaperTik