Recognition and extraction of honorifics in chinese diachronic corpora

Abstract

Honorifics in this paper refer to names of official positions and titles of nobility or honor. They can be found in various written records in different periods and have great historical significance. This paper introduces a machine learning system to recognize the honorifics in diachronic corpora. A tagged corpus of four classic novels written in the Ming and Qing dynasties is used to train the system. The system is then used to automatically recognize and extract the honorifics in pre-Qin classics, Tang-dynasty poems, and modern Chinese news. Experimental results show that the system can achieve relatively good results in recognizing the honorifics in the pre-Qin classics and Tang-dynasty poems. This work is an attempt to improve the performance of automatic recognition of honorifics in diachronic corpora. The system can be a helpful tool in the studies on the evolution of honorifics throughout Chinese history.