以通勤出行者为研究对象,应用改进K2算法和贝叶斯参数估计方法,以模块化的建模思想,构造了分析通勤出行方式选择和出行链模式安排及其相互作用的贝叶斯网络模型.以互信息指标度量节点间相互依赖关系的强弱,完成网络的修剪.以修剪后的网络为基础,应用敏感性分析讨论了在出行者及其家庭的社会经济属性、活动和出行属性影响下的出行方式和出行链模式安排的变化,及其相互影响.本文的研究为全面分析活动-出行选择行为及其影响因素间的互动响应关系提供了新的思路.%In this paper, a Bayesian network was developed to investigate the mode choice decision and trip chaining behavior of commuters as two interrelated modules. The model was based on the K2 algorithm and Bayesian estimation method. The original model was pruned based on mutual information of finding variables for travel mode and trip chaining choice variables. A detailed sensitivity analysis report of the pruned network was provided for a quantitative evaluation of the influence of significant finding variables on the two travel behavior choices. The results provide useful insights into the effects of sociodemographics, activity and travel characteristics on commuters' travel model and trip chaining propensity. This study also provides an important analysis tool for a comprehensive research of activity-travel pattern based on the influence factors and their interrelations mechanism.
展开▼
