基于数据挖掘技术的常规公交服务水平评价体系研究

发布时间：2018-03-16 23:28

本文选题：数据挖掘　切入点：常规公交服务水平　出处：《西南交通大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着城市规模的急剧扩大与人口数量的快速增长,常规公交已成为城市居民日常出行的主要方式,如何评价与提升常规公交服务水平已成为城市交通发展的主要目标之一。为了将评价结果有针对性地应用到改进措施中,论文以一条公交线路为研究对象,采用数据挖掘技术对常规公交服务水平进行评价。论文首先介绍了六种公交数据调查与获取的常用方法,并对各种方法的优劣与适用性进行分析。结合公交服务水平评价的要求与公交数据样本的特点,论文明确了公交数据挖掘的目标与工作流程,重点介绍分类分析与聚类分析两种任务与相应的工具模型,并对其进行实例探讨,决定采用决策树模型对公交服务水平评价体系进行分类挖掘研究。接着,论文以一条公交线路为研究对象,选取线路、车辆与营运三个层次共9个指标构建常规公交服务水平评价体系,并利用层次分析法对所选指标进行合理性检验。基于信息传递理论与决策树基本理论,论文提出了常规公交服务水平评价数据挖掘模型(DRBSE模型),并详细阐述了该模型的挖掘流程：建立决策树、决策树剪枝与提取模型规则。在建立决策树时,对指标的顺序选择进行探讨,使用信息增益率代替信息增益计算该属性的信息传递值,同时选择后剪枝法对决策树进行剪枝,并按照模型生成的枝叶结构进行规则提取。论文的最后以四川省207条公交线路为样本实例,以SPSS-Clementine为软件操作平台,利用DRBSE模型对常规公交服务水平评价体系进行建模,运行结果按照数据概率比高于70%的标准整理提取10条树形规则与指标重要度排序,经分析得出不同环境下提升常规公交服务水平的相应建议,验证了数据挖掘技术对常规公交服务水平评价的优越性。
[Abstract]:With the rapid expansion of urban scale and the rapid growth of population, bus routine has become the main way of daily travel for urban residents. How to evaluate and improve the service level of conventional public transport has become one of the main goals of urban traffic development. In order to apply the evaluation results to the improvement measures, the paper takes a bus route as the research object. Data mining technology is used to evaluate the service level of conventional public transport. Firstly, six common methods of public transportation data investigation and acquisition are introduced in this paper. Combined with the requirements of bus service level evaluation and the characteristics of bus data samples, the paper clarifies the goal and workflow of bus data mining. This paper mainly introduces two kinds of tasks and corresponding tool models of classification analysis and cluster analysis, and discusses them by examples, and decides to use the decision tree model to study the classification and mining of the evaluation system of bus service level. This paper takes one bus route as the research object, selects the line, the vehicle and the operation three levels altogether 9 indexes constructs the conventional public transport service level appraisal system. Based on the theory of information transfer and the basic theory of decision tree, the rationality of the selected index is tested by AHP. In this paper, a DRBSE model for the evaluation of bus service level is proposed, and the mining process of the model is described in detail: establishing decision tree, pruning decision tree and extracting model rules. The sequential selection of the index is discussed. The information gain rate is used instead of the information gain to calculate the information transfer value of the attribute. At the same time, the decision tree is pruned by selecting the pruning method. At the end of this paper, 207 bus routes in Sichuan Province are taken as sample examples, SPSS-Clementine as software operating platform, and DRBSE model is used to model the evaluation system of bus service level. According to the standard of data probability ratio higher than 70%, the operation results extract 10 tree rules and index importance ranking, and through the analysis of the corresponding suggestions to improve the general bus service level in different environments. The superiority of data mining technology in the evaluation of bus service level is verified.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP311.13;U491.17

【参考文献】