A new approach to Chinese person names recognition based on part-of-speech detecting

Automatic Chinese named entity recognition is an important task of Chinese information processing. This paper presents a new approach smartly combining statistical technique, heuristic rules and part-of-speech to identify Chinese person names. This algorithm features in boundary affirmation of person names by detecting candidate characters´ part-of-speech and in comparing the relative probability of would-be-word character. This work is partly funded by Chinese Natural Science Foundation under the grant number 60173060.