基于潜在类别模型的网约车用户市场细分

Market Segmentation of Taxi-hailing Users Based on Latent Class Model

为深入理解网约车用户出行需求,使用潜在类别模型对网约车用户进行市场细分。选取网约车使用频率、出行特征相关属性以及个人社会经济属性等影响因素对网约车选择行为建模,使用效率实验设计法生成意向偏好问卷,在成都市开展了实地调查。模型标定结果表明:在通勤和非通勤场景下,将网约车用户细分为四类最合适,两个场景下的市场结构比例均约为5:2:2:1。性别、月收入、时间、费用、网约车使用频率、私家车使用权等是对网约车用户进行市场细分的重要依据。

In order to understand travel demand of taxi-hailing users in a deep way,latent class model was used to study market segmentation of taxi-hailing users.Use frequency of taxi-hailing car,travel related attributes and individual social demo-graphic characteristics were chosen as the impact factors for modelling the taxihailing car choice behavior.Efficient experimental design was introduced to generate a stated preference questionnaire,and the field survey was conducted in Chengdu.Under the commuting and non-commuting scenarios,it is most appropriate to subdivide the taxi-hailing users into four classes,and the market share in both scenarios is about 5:2:2:1.Gender,monthly income,time,cost,frequency of taking taxi-hailing car,and ownership of private cars are also important basis for market segmentation of taxi-hailing users.