一种基于中文地理分词的动态交通信息数据模型研究

发布时间：2018-10-04 23:33

【摘要】：交通问题是目前城市发展中需要面临的一项重大问题,因此国家正在大力发展智慧城市,建设智能交通系统,以加快城市公交优化建设,提高大众交通信息服务能力。然而随着交通信息化的迅速发展,交通数据类型越来越繁多,数据量也日渐增大。而动态交通综合数据的分析和处理也面临着严峻的挑战。面对交通数据的多样性和复杂性,更加高效和稳定的数据组织管理方式才能有效地组织和管理不同类型的交通数据,同时,需要探索与研究数据的组织管理模型,思考现实交通数据问题的解决方案。即在数据的组织管理方面需要不断开拓时空数据模型研究,进一步提高数据存储与管理效率,注重海量数据的实时处理和深度挖掘,从而向社会和公众提供更加快速、准确、有效的交通信息服务以解决大众日常出行的公交慢、停车难、步行绕等问题。本文基于复杂多样的动态交通信息,围绕被动交通信息服务与主动交通信息搜索研究了信息检索的相关问题,并在探究中文地理分词方法的基础上研究了一种动态综合交通信息的数据模型并进行了应用探索。主要完成的工作有:(1)研究了中文地理分词方法,并在交通地理信息领域进行应用。为提高交通信息搜索时的服务响应速度,首先提出了交通专题信息快速切分方法;其次,实现了对RMM算法的改进,即实现了一种交通专题词优先的分词算法。该算法能够有效避免对专业词的歧义切分,结合结构优化的交通信息词典,能够提高对交通信息检索语句的分词精度。(2)面对上述算法的需要,并借鉴本体在语义表达上的层次特性设计了交通信息词典的存储结构,然后利用相关动态交通专题数据建立了交通专题信息词库。(3)基于本体方法建立动态交通信息综合数据模型。按照本体建模理论,并且结合地理本体的研究成果完成针对多专题动态交通信息的交通信息本体模型建立。(4)对模型的应用验证。以西城区TOCC系统为例,按照本体的模型结构将各种动态信息进行类别层次划分和关联整理。然后,利用分词算法对用户搜索语句做交通关键词提取,最终依据交通本体的推理查询,提供了对多类型动态交通信息的一体化搜索和多方位展示。结果表明,该模型对多源动态交通数据分析与检索具有一定的实际意义。研究成果也可以为动态交通信息处理提供技术参考。最后对基于本体知识的动态综合交通信息的进一步挖掘和应用进行了探索并做出展望。
[Abstract]:Traffic problem is an important problem in urban development at present. Therefore, the country is vigorously developing intelligent city and building intelligent transportation system, in order to speed up the construction of urban transit optimization and improve the capacity of mass transit information service. However, with the rapid development of traffic information, traffic data types are more and more, and the amount of data is increasing day by day. The analysis and processing of dynamic traffic data is also facing a severe challenge. Faced with the diversity and complexity of traffic data, more efficient and stable data organization and management methods can effectively organize and manage different types of traffic data. At the same time, it is necessary to explore and study the data organization and management model. Consider the solution of real traffic data problem. That is, in the aspect of data organization and management, it is necessary to develop spatio-temporal data model research, further improve the efficiency of data storage and management, pay attention to real-time processing and deep mining of massive data, so as to provide the society and the public with faster and more accurate. Effective traffic information service to solve the problem of slow public transportation, difficult parking, walking around and so on. Based on the complex and diverse dynamic traffic information, this paper studies the related problems of information retrieval around passive traffic information service and active traffic information search. On the basis of exploring the method of Chinese geographical word segmentation, this paper studies a data model of dynamically synthesizing traffic information and explores its application. The main works are as follows: (1) the Chinese geographic word segmentation method is studied and applied in the field of traffic geography information. In order to improve the service response speed of traffic information search, a fast segmentation method of traffic topic information is proposed firstly, and then the RMM algorithm is improved, that is, a word segmentation algorithm with traffic topic word first is implemented. The algorithm can effectively avoid ambiguous segmentation of professional words and improve the segmentation accuracy of traffic information retrieval statements by combining with the traffic information dictionary with optimized structure. (2) facing the need of the above algorithms, The storage structure of the traffic information dictionary is designed based on the hierarchical characteristics of ontology in semantic expression. Then, the lexicon of traffic topic information is established by using the relevant dynamic traffic thematic data. (3) the integrated data model of dynamic traffic information is established based on ontology method. According to ontology modeling theory and combined with the research results of geographical ontology, the traffic information ontology model for multi-topic dynamic traffic information is established. (4) the application of the model is verified. Taking Xicheng District TOCC system as an example, various dynamic information are classified and connected according to the model structure of ontology. Then, the segmentation algorithm is used to extract the traffic keywords from the user search sentences. Finally, according to the reasoning query of the traffic ontology, the integrated search and multi-directional display of the multi-type dynamic traffic information are provided. The results show that the model has some practical significance for the analysis and retrieval of multi-source dynamic traffic data. The research results can also provide a technical reference for dynamic traffic information processing. Finally, the further mining and application of dynamic integrated traffic information based on ontology knowledge are explored and prospected.
【学位授予单位】：北京建筑大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：P208

【相似文献】