Planning in-home activities in the ADAPTS activity-based model: a joint model of activity type and duration

The role of in-home activities in the process of planning and scheduling of individuals’ daily activities has been traditionally ignored because of two reasons: (i) in-home activities are not directly involved with trips; and (ii) scarcity of data sources that provide required details on planning and scheduling of these activities. However, considering the interchangeable nature of out-of-home and in-home activities, and the significant effects that they have on each other, we argue that failing to incorporate in-home activities may result in overestimating frequency and duration of out-of-home activities, which may lead to inconsistent and unrealistic activity schedules. Recently, we have upgraded the ADAPTS activity-based model to account for planning and scheduling of both in-home and out-of-home activities. This paper aims to enhance the in-home activity planning module by modelling the type and duration of the in-home activities in a joint structure. To achieve this goal, using the American Time Use Survey data, we estimate joint discrete-continuous models, in which activity type and activity duration are estimated by a multinomial logit and a log-linear regression model, respectively. The joint structure of these two models is established using copula approach that captures the unobserved shared factors affecting the two activity attributes. The results indicate that the estimated joint models significantly outperform the independent models in terms of goodness-of-fit and prediction accuracy.