基于多项Logit模型的交通出行方式预测

Prediction of Travel Mode Based on Multinomial Logit Model

为进一步量化分析不同因素对居民出行方式选择的影响,通过对合肥市居民出行进行调查,从交通规划的角度出发,构建了标准多项Logit交通方式划分预测模型,并采用极大似然法对模型参数进行估计。建模结果表明:1)模型的拟合优度和IIA假设检验均满足要求;2)女性驾驶员、私家车、驾照、高峰小时出行、出行目的、年收入、年龄7种因素对居民出行方式的选择有显著影响,研究结果可以为交通管理部门提供理论参考。

The impact of different factors on the choice of residents’ travel mode was further quantitatively analyzed.By investigating the travel of residents in Hefei, a Standard Multinomial Logit(MNL) transportation mode prediction model was built from the perspective of transportation planning and the model parameters was measured by the max-imum likelihood method. The results show that: 1)both the goodness of fit of the model and the Independence of ir relevant alternatives(IIA) hypothesis test met the requirements;2) 7 factors of sex,private car, driving license,travel in rush hour, travel purposes, income, age have significant impact on the choice of residents’ travel mode.The research results can provide theoretical reference for traffic management departments.