Application of adaptive Chinese segmentation in chat robot

Developing with the growing information society, people want to communicate with the computer by natural language and chat robot is such an interactive system to communicate with human beings. The key is to understand that chat robot technology means the speaker´s statement which the basic structure is the word, so the word is a critical foundation in Chinese information processing. The methods through word-based statistical learning are currently the mainstream in Chinese word practice, due to the limitation of its training corpus, resulting in poor adaptive capabilities. This paper proposes an adaptive Chinese words´ method which is a hybrid model combining rules and language It is showed that the adaptive ability of this hybrid approach has been improved significantly, the segmentation accuracy rate above 97%, reducing the error rate of word segmentation caused by ambiguity effectively.