基于C-F模型的中文地址行政区划辨识方法  

Chinese address administrative divisions identification method based on C-F model

作　　者：李晓林[1,2] 张懿 周华兵[1,2] 李霖[3] LI Xiao-lin 1,2 ,ZHANG Yi 1,2 ,ZHOU Hua-bing 1,2 ,LI Lin 3(1.Department of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan 430205,China;2.Hubei Provincial Key Laboratory of Intelligent Robot,Wuhan Institute of Technology,Wuhan 430205,China;3.Department of Resources and Environmental Science,Wuhan University,Wuhan 430079,Chin)

机构地区：[1]武汉工程大学计算机科学与工程学院,湖北武汉430205 [2]武汉工程大学智能机器人湖北省重点实验室,湖北武汉430205 [3]武汉大学资源与环境科学学院,湖北武汉430079

出　　处：《计算机工程与设计》2018年第7期2048-2053,共6页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2013AA12A202);测绘地理信息公益性行业科研专项基金项目(201412014);湖北省自然科学基金项目(2013CFA123);武汉工程大学第八届研究生创新基金项目(CX2016065)

摘　　要：为有效解决互联网中表示地理位置的中文地址行政区划信息模糊、缺失和歧义的问题,结合不确定性推理技术,提出一种基于C-F模型的区划可信度计算方法。依据C-F模型,构造可信度计算方法,对通过移动窗口最大匹配算法得到的行政区划结论进行不确定性推理计算,选取可信度相对最大的区划结论作为辨识结果。实验结果表明,该方法可通过识别与完善中文地址行政区划,提高地址辨识准确性,将实验结果与其它方法进行比较分析,验证了该方法的可行性和有效性。To deal with the problem that Chinese address administrative division is fuzzy,missing and ambiguity,the calculation method based on C-F model with uncertainty reasoning technology was proposed.The credibility of administrative divisions which were acquired using the maximum moving window matching algorithm with making use of credibility factor calculation method was calculated,and the division information with the highest credibility factor in the result was selected to be the result of identification.Experimental results show that the proposed method can improve the accuracy of address identification by identifying and perfecting the administrative division of Chinese address.Compared to other methods,this method is feasible and effective.

关 键 词：C-F模型 行政区划 移动窗口匹配 不确定性推理 可信度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]