Predicting Individual Hiking Trail Intensity Using Statistical Learning
Digital Commons - USU (Utah State University) · 09.08.2024
DOI: 10.26076/2a55-c974 ↗Kurz gefasst
Fachartikel in der Zeitschrift Digital Commons - USU (Utah State University), erschienen am 09.08.2024. Verfasst von Kelci B. Hannan.
Abstract (Originalsprache)
Hiking is an increasingly popular choice for people wanting to engage in recreation and physical activity. It not only offers the many benefits of exercise but also the opportunity to explore nature and to socialize. To ensure that new hikers are prepared to hike and remain safe while on the trail, knowledge of a hiking trail's difficulty is important. Trail difficulty information is sometimes available at the trailhead if the trail is well-trafficked and maintained but is nearly always available online or through a hiking-specific mobile application, such as AllTrails. Both trailhead and online sources provide only one difficulty rating for all users regardless of personal characteristics, such as age and fitness level, which can lead to a misalignment of expectations and actual hiking experience. This project's overall goal was to attempt to generate personalized predictions of hiking trail difficulty and assess the effectiveness and utility of this approach. During a series of investigations, individuals hiked the Wind Caves trail in Logan, Utah while wearing a fitness watch that recorded heart rate, hiking speed, and location. Participants recorded how hard they felt they were exercising while hiking using a metric called Rating of Perceived Exertion. Before hiking, each participant answered questions about their fitness level, response to painful situations, and hiking experience. This information combined with statistical models was used to predict each individual's hiking intensity on the Wind Caves trail. These models were then evaluated for accuracy to assess how they might be used to improve estimates of hiking trail intensity prior to beginning a hike. We also explored which variables most improved hiking intensity prediction and how different perceived exertion scales impacted model performance. The results of this project suggest that personalized predictions of hiking difficulty may be feasible with refinement of the approaches used here, such as imposing one hiking speed or focusing on a sub-group of hikers, and lay a foundation for future research into applying statistical models to this question. Finally, we offer areas for improvement in future studies examining personalized trail intensity predictions.
Schlagwörter
Zitieren
APA
Hannan, K. B. (2024). Predicting Individual Hiking Trail Intensity Using Statistical Learning. Digital Commons - USU (Utah State University). https://doi.org/10.26076/2a55-c974
Vancouver
Hannan KB. Predicting Individual Hiking Trail Intensity Using Statistical Learning. Digital Commons - USU (Utah State University). 2024. doi:10.26076/2a55-c974
BibTeX
@article{Hannan2024,
author = {Hannan, Kelci B.},
title = {Predicting Individual Hiking Trail Intensity Using Statistical Learning},
journal = {Digital Commons - USU (Utah State University)},
year = {2024},
doi = {10.26076/2a55-c974}
}
EUVIMED