Word Segmentation Model Of The Ancient Chinese Based On Root Word Algorithm

Word segmentation of Chinese is an important subject of computational linguistics, and it is also the first step of some work,such as Machine Translation, Text Classification, Information Retrieval, Speech Synthesis. At the same time, it is also a bottle-neck problem of NLP. Now, the word segmentation technology of modern Chinese has developed rapidly, the precision of some segmentation systems are up to 95%. However there are few reports on word segmentation of the ancient Chinese.In this paper, I design a Word Segmentation Model of the Ancient Chinese Based on Root Word Algorithm, which is based on the features of the texts of the ancient Chinese words, the knowledge of the ancient Chinese linguistics, as well as the statistical information of 80% of the ancient Chinese words are single-character word. Then I give a formalized description of this algorithm, and compare it with the wildly-used Maximum Matching Method. Finally I realized a test program based on this algorithm through VC++.The design of segmentation dictionary based on corpus affects the precision and speed of the word segmentation. This paper proposes some problems aimed at the design of the segmentation dictionary of the ancient Chinese after introduce the development of the corpus in China. The ambiguity resolution of the experimental word segmentation result is also an important part of the word segmentation model. Then I introduce the kind and cause of the ambiguity of the ancient Chinese word segmentation. After summarize the characteristics of syntax of the ancient Chinese, I propose a strategy of the ambiguity resolution based on the sentence pattern of the ancient Chinese.