Corpus based Semi-Automatic Extraction of Persian Compound Verbs and their Relations

Nowadays, Wordnet is used in natural language processing as one of the major linguistic resources. Having such a resource for Persian language helps researchers in computational linguistics and natural language processing fields to develop more accurate systems with higher performances. In this research, we propose a model for semi-automatic construction of Persian wordnet of verbs. Compound verbs are a very productive structure in Persian and number of compound verbs is much greater than simple verbs in this language This research is aimed at finding the structure of Persian compound verbs and the relations between verb components. The main idea behind developing this system is using the wordnet of other POS categories (here means noun and adjective) to extract Persian compound verbs, their synsets and their relations. This paper focuses on three main tasks: 1.extracting compound verbs 2.extracting verbal synsets and 3.extracting the relations among verbal synsets such as hypernymy, antonymy and cause.