Распределімдік ұқсастық жүйесі

Most existing systems for automatically extracting lexical-semantic resources ne- glect multi-word expressions (MWEs), even though approximately 30% of gold-standard thesauri entries are MWEs. We present a distributional similarity system that …...