摘要
鉴于MNL模型的非相关选择方案独立性易引起预测误差,同时考虑到西藏地区在全国旅游市场的特殊地位,选用NL模型对拉萨游客出行行为进行研究,并引入游客主观意向选择修正系数α,选取景点和出行方式分别作为水平2和水平1的选择肢,以出行费用、行程时间、年龄、收入、是否有私家车作为效用变量,在水平1的基础上提出了两种交通方式的划分方法,分别构建了以游客出行意向和交通特性为出发点的NL模型,利用拉萨游客出行调查数据对模型进行参数标定,检验结果表明以游客出行意向为出发点的划分方式更符合实际情况,引入主观意向修正系数后,预测的相对误差由11.82%降至9.99%。
The independence of MNL model is easy to cause the prediction error. Taking into account the special status of Tibet in the national tourism market, the NL model is choosed to study the travel behavior of Lhasa tourists, and the subjective intention selection correction factor α of tourists is introduced. Attractions and travel mode are selected as alternative level 1 and level 2, travel costs, travel time, age, income, have private cars or not are selected as utility variables. Based on the level of a division on the proposed two methods of transportation , it was constructed with the intention of tourists travel and transportation characteristics as the starting point of NL model. The parameters of the model are calibrated by the data of Lhasa tourists travel survey, test results show that the division method according to tourists travel intention is more accordant with actual situation, the relative error of the prediction is reduced from 11.82% to 9.99% considering the correction coefficient of subjective intention.
出处
《现代交通技术》
2018年第1期65-70,共6页
Modern Transportation Technology
基金
国家科技支撑计划(项目编号:2014BAG05B00)
关键词
交通工程
选择行为
NL模型
主观意向
随机效用
traffic engineering
choice behavior
NL model
subjective intention
stochastic utility