基于Logit模型的共享汽车出行影响因素分析

为定量分析影响使用共享汽车出行的主要因素,以543份有效调查问卷收集的数据为基础,从个体特征、出行特征和出行者对共享汽车服务水平的主观感知3个方面开展研究,分别运用有序Logit模型和多项Logit模型建立共享汽车出行频率预测模型,并对比预测精度。结果表明,职业类型、对共享汽车舒适性、便利性、经济性和时耗性的主观感知5个变量与共享汽车出行频率显著相关。有序Logit模型和多项Logit模型的平均预测准确率分别为78.41%和92.40%,多项Logit模型的预测效果较好。

Evaluation of Car Sharing Factors Based on the Logit model

In order to analyze quantitatively the main factors influencing the choice of car sharing travel, based on data collected by 543 valid questionnaires, according to three aspects including individual characteristics, travel characteristics and subjective perception of car sharing service level, this paper respectively establishes ordered Logit model and multinomial Logit model to predict car-sharing travel frequency and compares forecast accuracy of two both models. The results show that the five variables including the type of occupation and subjective perception of comfort, convenience, economy and time consumption of car sharing are significantly related to the travel frequency of car sharing. The average prediction accuracy of the ordered Logit model and multinomial Logit model is respectively 78.41% and 92.40%, so the prediction effect of multinomial Logit model is better.