基于潜变量的驾驶员路径选择行为分析

Drivers' Route Choice Behavior Analysis Based on Latent Variables

为更真实地还原驾驶员的路径决策过程,构建了基于潜变量的路径选择模型以增强模型的解释性。首先采用结构方程模型识别行为中存在的潜变量;并通过离散选择模型分析潜变量对选择行为的影响。最后以广州市驾驶员为例进行实证分析。结果表明:在驾驶员的态度与行为中存在4个潜变量因子;且各潜变量与驾驶员的社会经济属性有关。驾驶员的路网熟悉度与路径探索意识对其路径选择行为有较大影响;考虑潜变量的模型有效提高了模型的拟合度。

In order to get a realistic description of drivers＇ route choice behaviors, latent variables were intro- duced into the traditional route choice model. A structural equitation model is established to recognize latent factors in driving behaviors, then these latent variables are incorporated into the discrete choice model. A case study of Guangzhou is conducted and results revealed that： Four latent factors are found in the case and their values are varied among drivers. The road network familiarity and the route exploring consciousness are two important latent variables to their choice behaviors. Considering latent variables enriches the comprehension of the model.