基于贝叶斯网络的城市常规公交服务满意度分析模型

发布时间：2018-12-21 19:46

【摘要】：为揭示城市公交服务指标之间的潜在相关性和分析各指标对常规公交整体满意度的影响度,以609份南京市常规公交服务调查问卷为数据基础,利用GTT算法和MLE算法,建立基于贝叶斯网络公交服务满意度分析模型,挖掘常规公交服务指标与乘客总体满意度之间的网络关系.模型检验结果显示,贝叶斯网络具有较高的适用性和有效性.由置信度传播算法获得的推理结果得到的7项主要影响因素中,候车时间的负效应-31.4%最为显著,车内卫生的正效应最为明显,高达33.7%,行车准点性和换乘便捷性各自具有较高的提升空间且正负影响域较大,分别为-26.6%~23.1%和-20.9%~27.3%,车内有无座椅、候车环境和空调的配备具有较高的绝对影响域,均高达55%.
[Abstract]:In order to reveal the potential correlation between urban public transport service indicators and analyze the influence of each index on the overall satisfaction degree of bus routine, based on 609 questionnaires of routine bus service in Nanjing, the GTT algorithm and MLE algorithm are used. Based on Bayesian network bus service satisfaction analysis model is established to explore the network relationship between conventional bus service indicators and passenger overall satisfaction. The results of model test show that Bayesian network has high applicability and effectiveness. The negative effect of waiting time was the most significant, and the positive effect of in-vehicle hygiene was the most obvious, as high as 33.7%, among the seven main influencing factors obtained from the reasoning results obtained by the confidence propagation algorithm, the negative effect of waiting time was 31.4%, and the negative effect of waiting time was 31.4%. The traffic punctuality and transfer convenience have higher lifting space and positive and negative influence areas, which are -26.623.1% and -20.9% 27.3, respectively. There are no seats in the car, Waiting environment and air-conditioning have a higher absolute impact area, up to 55.
【作者单位】：东南大学城市智能交通江苏省重点实验室;东南大学现代城市交通技术江苏高效协同创新中心;东南大学交通学院;
【基金】：国家自然科学基金资助项目(51338003,51378120)
【分类号】：U491.17

【相似文献】