Statistical model added linguistic rules for Mongolian lexical analysis
Lili Zhao · Jisuanji gongcheng yu sheji · 2011
Mongolian lexical analysis is the first step in Mongolian information processing.Statistical model and Mongolian language rules are researched,a method of integrating statistics and rules technology is presented which can solve the Mongolian word segmentation and POS tagging at the same time.This method uses a statistical model as the basic framework to describe the system with tree structure the solution space,then added the Mongolian linguistic rules and the stem,affix dictionary.The results show that the system which joint word segmentation with POS tagging can reach 95.1% in word-level accuracy,and 94.6% in stem / affix-level accuracy.The ex-perimental results have been improved compared with other lexical analysis system.