| dc.description.abstract | Cycling behavioral research is increasingly conducted by means of GPS data. The presence of
these data-sets allow for large-scale investigation of complicated travel behavioral aspects such as
cycling motives and enables one to enrich raw GPS data-sets with these attributes based on
contextual information. Currently, both the differences in cycling behavioral between cyclists with
different motives and the extent up to which these differences can be used to estimate cycling
motives for raw GPS tracks have received little attention. Even though more insights on these
topics can provide useful insights for policymakers and can stimulate travel behavior research by
enabling others to enhance their GPS tracks with more accurate cycling motive attribute data.
This research tries to tackle both these problems by establishing cycling behavioral profiles based
on trip, route and origin-destination behavioral characteristics and subsequently using the
differences in these profiles to estimate cycling motives by means of machine learning. In addition
to that, multiple machine learning algorithms are assessed to determine the most suitable The
results show that there are significant differences in cycling behavioral profiles between motives.
Trip, route and origin-destination behavioral characteristics all outperform a standard model for
estimating cycling motives, with a combined model including all behavioral characteristics scoring
highest (74.0% accuracy versus 51.4% standard model accuracy). Furthermore the results indicate
that Random Forest and Gradient Boosting are among the most suitable algorithms for this
purpose. Finally, recommendations and potential improvements are provided for future research on
cycling behavior and motive estimation. | |
