自然语言处理视角下分散化政策情报信息量化研究——以1986-2018年旅游交通政策演进为例  

The Quantitative Research of Decentralized Policy Intelligence Information from the Perspective of Natural Language Processing——Taking the Evolution of Tourism Transportation Policy from 1986 to 2018 as an Example

作　　者：魏宇 邢剑华[1,2] Wei Yu;Xing Jianhua(School of Economics and Management,Beijing Jiaotong University,Beijing 100044;Department of Tourism,Shanxi University,Datong 030006)

机构地区：[1]北京交通大学经济管理学院,北京100044 [2]山西大学旅游系,大同030006

出　　处：《情报杂志》2020年第8期16-23,121,共9页Journal of Intelligence

基　　金：国家自然科学基金项目“综合交通运输系统中风景道评估体系及模式”(编号:40971288)。

摘　　要：[目的/意义]情报信息的分散化分布规律是情报活动的基石,一些交叉性、前沿性政策类情报文本,呈现明显分散化特征,探寻此类情报信息量化研究方法,具有重要学术意义和现实紧迫性。[方法/过程]该文利用自然语言处理方法(Natural Language Processing,NLP),对旅游交通政策这一交叉领域的分散化政策演进进行量化分析。[结果/结论]研究验证了该方法对分散化政策类情报信息分析的适用性。结果表明:基于规则的命名实体识别对政策总量与结构的分析显示,发文总量逐年上升,不同时期政策结构有所差异。基于规则和统计模型的命名实体识别对政策制定主体的分析显示,政策颁发机构数量呈起伏上升趋势,多部门参与局面渐显。基于词性和句法分析的政策内容演进显示:“事故”始终是政策内容的语义核心词;语义聚焦主题词演进兼具继承性和发展性。[Purpose/Significance]The law of distribution of intelligence information is the cornerstone of intelligence activities.Some cross-cutting and frontier policy-type intelligence texts present obvious decentralized characteristics.It is of great academic significance and practical urgency to explore the research methods of intelligence information quantification.[Method/Process]Using Natural Language Processing(NLP),this paper conducts a quantitative analysis of the policy evolution in the cross-domain decentralization of tourism transportation policy.[Result/Conclusion]The research validates the applicability of NLP in the quantitative analysis of cross-cutting and decentralized policy.The results show that:Analysis of policy aggregate and structure using rule-based named entity recognition shows that:policy aggregate indicates an upward trend year by year,and the policy structure varies from time to time.Analysis of policy-making entities based on rules and statistical models of named entity recognition shows that:the number of agencies has a rising trend;the number of departments enacting policies at various stages has gradually increased.Analysis of the evolution of policy content based on part-of-speech and syntactic analysis shows that:"Accident"is always the semantic core word of the policy content;the evolution of semantic focus theme has both inheritance and development.

关 键 词：自然语言处理 情报信息 量化研究 分散化政策 旅游交通 

分 类 号：F590[经济管理—旅游管理]