A LATENT CLASS MODEL FOR A EXOGENOUS SEGMENTATION APPROACH

点击翻译按钮获取中文摘要

This paper proposes an exogenous segmentation approach to some choice behavior models. This approach jointly determines the number of market segments in the travel population, assigns individuals probabilistically to each segment, and develops a distinct choice model for each segment group. The author proposes a stable and effective estimation approach for the endogenous segmentation model that combines an Expectation-Maximization algorithm with initial value setting. The exogenous segmentation model and other commonly used models in the travel demand field to capture systematic heterogeneity are estimated using some choice behavior datasets. The results show that the endogenous segmentation model using latent class model fits the data better and provides meaningly more reasonable results compared to the other endogenous approaches.