In this study a sequential model of activity patterns is formulated that consists of time- and history-dependent models of activity choice. This analytical framework is used to identify time-of-day and history-dependent characteristics of activity choice by statictically testing a series of hypotheses. The results indicate that the simplest expression of the history of activity engagements is an adequate descriptor, and also that non-home-based activity choice is conditionally independent of the activities in the previous chains, given the activities pursued in the current trip chain. Interdependencies of activity types across trip chains are also characterized by estimated model coefficients. The results of the study indicate that the decisions associated with the entire activity pattern can be decomposed into interrelated activity choices whose conditional dependencies can be statistically evaluated.
Abstract