一类蒙文词切分方法构造及智能词切分系统设计  

Some Methods for Mongolian Word Segmentation and Its Intelligent System Design

作　　者：阿日木扎[1] 林民[1] 马占新[2] Arimuzha;LIN Min;MA Zhan-xin(College of Com puter Science and Technology,Inner Mongolia Normal University,Hohhot 010022,China;School of Economics and Management,Inner Mongolia University,Hohhot 010021,China)

机构地区：[1]内蒙古师范大学计算机科学与技术学院,内蒙古呼和浩特010022 [2]内蒙古大学经济管理学院,内蒙古呼和浩特,010021

出　　处：《内蒙古师范大学学报（自然科学汉文版）》2019年第5期393-399,共7页Journal of Inner Mongolia Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(70961005,71261017);内蒙古自治区蒙古语言文字科研资助项目(MW-YB-2016033,MW-YB-2018023)

摘　　要：实用高效的蒙古文词切分方法是实现蒙古文信息处理的前提和基础.研究提出从3种基本的蒙古文词切分方法,组合成12种蒙古文词切分方法.应用数据包络分析(DEA)方法,提出用于蒙古文词切分方法综合评价的数学模型和有效性计量公式,并围绕蒙古文切词方法的智能选择问题,提出基于国际标准编码的蒙古文词切分系统的设计方法和整体框架.A feasible and effective Mongolian word segmentation method is the premise and foundation to realize excellent Mongolian information processing.In this paper,12 kinds of Mongolian word segmentation methods were produced from recombination of 3 basic Mongolian word segmentation methods firstly.Then,a mathematical model and some measurement formulae for comprehensively evaluating the validity of Mongolian word segmentation methods were derived by using data envelopment analysis(DEA)method.Finally,we gave out a design method and its overall framework of Mongolian word segmentation system based on the international standard code in regard of effectively choosing a Mongolian word segmentation method.

关 键 词：蒙古文信息 国际编码 词切分 数据包络分析 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]