大数据背景下道路运输运营管理数据挖掘研究与应用

发布时间：2018-02-11 09:00

本文关键词： 道路运输数据挖掘聚类粗糙集关联规则　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,随着移动互联网、物联网、计算机技术等新兴科技的盛行,行业应用产生的信息呈现爆炸式发展的态势,“大数据”一词骤然成为各行各业炙手可热的新名词。道路运输行业也紧跟时代潮流,基于车联网和信息技术等新兴科技催化下的信息资源,具有规模性、高速性、复杂性、多样性等“大数据”特征。通过高效开发、利用道路运输信息资源的“加工”能力实现其“增值”作用,为行业管理提供决策支持,已是大势所趋。为此,本文依托道路运输中的核心信息系统进行数据挖掘,旨在能够助力行业运营管理。数据挖掘技术作为一门多领域、多学科的融合技术,在各行各业中应用广泛。本文在借鉴国内外学者关于数据挖掘理论、方法、行业应用等的研究成果基础上,通过深入探讨实施大数据分析的基础条件,即历史信息、资源共享、决策目标,明确了本文的研究主题。提出应用聚类分析、粗糙集、关联规则方法改善现有道路运输经营状况,并得到有效性验证。基于班线客运市场运营中现存的运输模式常年固定不变的问题,从客运线路、发车日期、发车时间三个角度进行数据挖掘。通过合理的数据预处理和分析指标的选择,采用Two Step聚类分析方法,即构建CF树将数据集预聚类为较小子集,再应用凝聚的层次聚类合并为最佳数量的类别。应用SPSS Modeler软件建立数据流模型并实现结果的可视化展示。最终得出旅客频繁出行的线路、时间的规律,为客运管理者开展以旅客为中心,按需定制、灵活发班的客运经营模式提供决策依据。基于卫星定位信息系统中的运输业户考核信息数据,结合粗糙集理论与关联规则方法,对数据进行预处理约简来提高生成规则的质量。从数据离散化、属性约简、规则构建三个方面设计算法,应用信息熵理论、遗传算法和Apriori方法实施数据挖掘。通过条件属性的约简明确影响考核结果的关键因素,并缩减规则构建的输入信息,以简洁明了、科学有效的决策规则来助力运输业户和行业管理者对卫星定位平台的监管。本文将数据挖掘中的主要方法应用于对信息系统数据的分析,优化了道路运输中客运和车辆监管方面的宏观管理和调控,为行业管理提供了的决策支持。
[Abstract]:In recent years, with the mobile Internet, networking, computer technology and other emerging technologies of popular industry application information generated by the explosive development of the situation, the big data word suddenly become the hottest new term in all walks of life. The road transport industry is also closely follow the trend of the times, car networking and information technology and other new technology under the catalysis of information resource based on the scale, speed, complexity, diversity and other characteristics of "big data". Through the efficient development and utilization of road transport information resources of the "process" to achieve their "value-added", to provide decision support for the industry management, is represent the general trend. Therefore, the core information system based on road transport the data mining, in order to operation and management of power industry. Data mining technology is a multi domain fusion technology, widely used in all walks of life. Referring to the domestic and foreign scholars on the theory of data mining, method, application based on research results, through the discussion on the implementation of the basic conditions of large data analysis, namely historical information, resource sharing, decision objective, clear the theme of this paper. We should use clustering analysis, rough set, Association rules to improve the existing methods the road transport operations, and validation. The class line passenger transport mode existing in the market operation of perennial fixed problem based on departure from passenger lines, date, data mining three point of departure time. Through reasonable data preprocessing and analysis index, using Two Step cluster analysis method, i.e. build the CF tree data set for a minimal set of pre clustering, hierarchical clustering and application of condensed into the optimal number of categories. The application of SPSS Modeler software to establish the data flow model and To achieve results visualization. Finally the passenger travel frequently line time rule, for managers to carry out passenger passenger as the center, on-demand, flexible decision-making based on class passenger business model. The satellite positioning information system in the transport industry examination information data based on the combination of theory and method of rough association rules set, preprocess data reduction to improve the quality of the production rule. From the data discretization, attribute reduction and rule construction in three aspects of the design algorithm, using information entropy theory, genetic algorithm and Apriori implementation method of data mining. The key factors affecting the assessment results clearly by reduction of condition attributes, the input information and reduction rules the construction, in a concise, effective scientific decision rules to help transport operators and industry managers to monitor the satellite positioning platform. This paper will in data mining The main method is applied to the analysis of information system data, optimized the macro management and regulation of passenger transport and vehicle supervision in road transportation, and provided decision support for industry management.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;F542

【相似文献】