Key points are not available for this paper at this time.
As the urban population continues growing and residents prioritise the green and healthy aspects of cities, urban parks are facing unprecedented pressure in terms of usage. Through physical activity survey and landscape morphology qualifying from urban parks in Nanjing, China, this study conducted predictive analysis on 5 activity classification patterns and 14 specific physical activities, by employing random forest and gradient-boosting tree classification models. Our findings indicate that classification based on gender and outdoor activity items exhibited superior average prediction accuracy. Moreover, among the 14 activities, better prediction results were obtained for activities such as rest, collective, female, children, and intimate activities. This research explores the potential for extending studies from the correlation between activity and environment to prediction, offering valuable insights to enhance the interactive analysis between park landscape design and environmental behaviour. Ultimately, it aims to promote the efficient utilisation of urban park environments.
Ma et al. (Fri,) studied this question.