作者: Xin Ye , Ram M. Pendyala
DOI: 10.1007/978-1-4419-0820-9_20
关键词:
摘要: This paper presents a probit-based discrete-continuous modeling methodology to analyze relationships between discrete and continuous choice dimensions often encountered in activity-travel behaviour research. The approach allows one adopt flexible multivariate normally distributed error covariance structure that overcomes the limitations associated with other methods rely on two-step limited information techniques or restrictive distributional transformations. detailed formulation of demonstrates its applicability through an analysis relationship timing (scheduling) duration household maintenance activities include shopping personal business activity episodes. A new non-nested test developed by authors is used compare alternative model structures identify nature joint activities. Models are estimated separately for commuter non-commuter samples drawn from 2000 Switzerland Microcensus Travel. Model estimation results show covariances significant models duration. Non-nested comparisons indicate specification where time-of-day affects offers statistically superior goodness-of-fit comparison choice. These findings lend credence notion adopted activity-based systems should be which schedule agenda drives time allocation