Word Segmentation Method For Oracle Bone Inscriptions Based On Dictionary And Syntactic Rules

According to Oracle Bone Inscriptions(OBI)'s own characteristics, proposed a combination of thesaurus-based word frequency, part of speech, syntactic rules segmentation algorithm. Firstly, the method obtained by the initial OBI dictionary segmentation result, and then using OBI's grammar rulse and word entry filter rules to get segmentation result; Finally, using the unknown words identification rules to check the result, the words which meet the unknown word's condition threshold will be added into the dictionary. The segmentation method can largely remove ambiguities and improve the identification of unknown words probability. Experimental results show that the method is consistent with the grammatical features of OBI, and it can obtain high accuracy and recall.