Automatic term extraction for chinese opera domain ontology

This paper aims at the research of automatic term extraction technology for Chinese opera domain ontology. It applies the hybrid method based on the association rule algorithm, C-value algorithm and TF-IDF algorithm to extract simple and complex terms automatically. Firstly, the association rule algorithm is applied to get candidate terms from domain corpus comprehensively. Secondly, the C-value algorithm is applied to filter out meaningless items and the TF-IDF algorithm is applied to filter out popular items with low correlation in Chinese opera domain effectively. The experiment shows that the association rule algorithm can guarantee that the recall of term extraction is quite high. Meanwhile, the C-value and TF-IDF algorithm can improve the precision of term extraction.