Intershopping duration: an analysis using multiweek data

Bhat, Chandra R.; Frusti, Teresa; Zhao, Huimin; Schönfelder, Stefan; Axhausen, Kay W.

doi:10.1016/s0191-2615(02)00093-0

Cited by 67 publications

(36 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of the empirical example provided earlier, the total number of systematic utility component computations per likelihood function iteration is of the order of 2 × 10 3 × Q × N in our approach, which is substantially lower than the 10 6 × Q × N computation in the full enumeration approach. The latent structure approach used here has been applied in the context of latent segmentation, where decision-makers are probabilistically assigned to each segment based on their characteristics, and a separate behavioral structure corresponding to systematic sensitivity variations to exogenous variables is estimated for each latent segment (26)(27)(28)(29). However, the novelty in the current paper is that the latent structure is being applied not to accommodate differential systematic variable sensitivity across individuals, but to accommodate different decision-making structures (i.e., whether a household first decides on residential location and then on workplace location, or vice versa).…”

Section: Introductionmentioning

confidence: 99%

Modeling Interdependence in Household Residence and Workplace Choices

Waddell

Bhat

Eluru

et al. 2007

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Models of residential and workplace location choice prevalent in the literature often assume that one choice dimension is exogenous to the other. In our view, a broad and uniform assumption that one choice dimension is exogenous and influences the other is too strong to use as the foundation for current behavioral research or applied policy analysis. We seek to examine the interdependence of residence and workplace choices and to develop a novel approach to modeling these choice dependencies. Two problems related to such joint modeling efforts are addressed in this paper. First, through a latent market segment modeling approach, the paper offers a methodology for accommodating different sequential decision-making processes that may be present in the population, i.e., residential location may be chosen first and influence workplace location for one segment and vice versa. Second, the modeling approach offers a means of overcoming the exploding choice set problem when attempting to model multidimensional choice phenomena. The overall aim of the work is to model the structure of the interdependency between the choices that a household makes with respect to residence location, and the workplace choices of the workers in the household, in the context of an integrated activity location and travel forecasting framework. This paper presents a joint model of residence location and workplace using an activity-based travel survey collected in the Puget Sound region of Washington in 1999, using a novel adaptation of recent methods for incorporating latent market segmentation within discrete choice models.Keywords: Residential location, workplace location, latent segmentation, discrete choice, behavioral decision-making, multi-dimensional choices Waddell, Bhat, Eluru, Wang, and Pendyala 3 INTRODUCTION Traditional urban economic theory and the vast majority of empirical work in urban modeling treats the residence location choice as conditional on an exogenous choice of workplace. Some empirical work, on the other hand, including the workplace destination choice models embedded in metropolitan transportation models, assume that residence location is predetermined when predicting the workplace location choice. All of this work treats the residence and workplace choices as though one is exogenous and influences the other. More importantly, such endeavors assume that the same conditional choice process applies to the entire population under study. The assumption that one choice dimension (i.e., either residence or workplace choice) is exogenous and influences the other for the entire population without any regard for the possibility that a reverse choice process may be in vogue for at least a segment of the population is, in our view, too strong to use as the foundation for current behavioral research or applied policy analysis. We seek to examine the interdependence of residence and workplace choices and to develop a novel approach to modeling these choice dependencies that overcomes two inter-related challenges. First, ...

show abstract

Section: Introductionmentioning

confidence: 99%

Modeling Interdependence in Household Residence and Workplace Choices

Waddell

Bhat

Eluru

et al. 2007

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

show abstract

“…[12] Bhat et al (2004) model the shopping activity generation indirectly by modelling the inter-shopping duration distribution considering the individual's affinity or dislike for shopping activities. [13] Bhat and Misra (1999) and Bhat (1998) model shopping activities combined with other out-of home activities. Although some of this research recognizes the household influence on individuals' shopping activity, [14,12] they consider the household effect either directly by incorporating household level socio-economic variables or propose very comprehensive modelling approaches that are often not supported by the data available from activity diary surveys.…”

Section: Modelling Shopping Activitiesmentioning

confidence: 99%

Modeling Individuals' Frequency and Time Allocation Behavior for Shopping Activities Considering Household-Level Random Effects

Habib¹,

Miller

2006

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

The paper describes a comprehensive frequency and time allocation modelling system for shopping activities. The modelling system is person-based but explicitly considers fixed and random household effects. It has three components: a weekly shopping frequency model, a daily shopping frequency model, and a time allocation model for individual shopping episodes. The frequency models consider activity generation as a latent response, the propensity to participate in shopping activities. This latent response is modelled using an ordinal response model. Both the weekly and daily frequency models are multilevel ordinal logit models, where the household is the highest level and the individual person is the lowest level. The multilevel ordinal logit models incorporate household level influences on an individual's shopping behaviour in term of fixed effects and a random intercept. The time allocation models are hazard duration models that consider household-level random heterogeneity. The entire modelling system is sequential from the weekly frequency component to the time allocation component. The outputs of the earlier components enter as inputs to the later components: weekly frequency is the input to the daily frequency model; weekly and daily frequencies are input to the time allocation model. The system is designed to overcome the deficiencies of household-level information in the personallevel activity diary data that are used to develop the model. Khandker M. Nurul Habib and Eric J. Miller

show abstract

“…Over longer time horizons these rhythms are observable also for other activity types. See for example Bhat, Frusti, Zhao, Schönfelder and Axhausen, 2004 (case of shopping), or Schönfelder and Axhausen, 2001 for all activity types.…”

Section: Rescheduling Executionmentioning

confidence: 99%