Abstract：The increase in non-commuting trips is an important indicator of a city’s high-quality development, especially for the
innovative population with high income but limited disposable time. They have a more urgent need for service facilities
that meet their requirements for high-quality and diverse non-commuting activities. This study, taking Hefei, one of rapidly
developing innovative cities in the interior, as an example, based on the perspective of activity types, explores the influencing
factors of residents’ non-commuting activity decisions by analyzing the characteristics of innovative people when they engage
in different types of non-commuting activities and using the random forest model and SHAP model. It also investigates
the mechanism of the impact of travel characteristics, built environment, and socio-economic attributes on non-commuting
activity decisions. The research findings are as follows: (1) The non-commuting activities of the innovative population can
be divided into two categories: family care and leisure time, with the latter being the main type. Their activity characteristics,
population characteristics, and built environment are all different. (2) The travel characteristics of different innovative talents
when they engage in non-commuting activities have the greatest impact on their decision-making for non-commuting activity
types, followed by the built environment of their residence, and socio-economic attributes have the least significant impact. (3)
Specifically, among the built environment variables, the number of significant factors is the largest, and there are both linear and
non-linear relationships with non-commuting activity decisions. The number of travel characteristic variables is relatively small,
but their relative importance is the highest, and they are mainly non-linear. Socio-economic attributes are mainly linear and their
importance rankings are all at the back. The research results will deepen the theoretical understanding of the non-commuting
activities of the innovative population and reveal the non-linear mechanism of variables. In practice, the research will provide
planning guidance and policy references for helping the innovative population improve their quality of life and promoting the
innovative development of inland cities.
