Interpretable Modeling of Physical Activity, Sleep, and Sedentary Behavior from High-Frequency Wearable Data

Xuefen Zhong; Siqi Zhou; Jingying Guo

doi:10.63944/1gnmqw74

Interpretable Modeling of Physical Activity, Sleep, and Sedentary Behavior from High-Frequency Wearable Data

Xuefen Zhong (作者)

School of Mathematics, Guangdong University of Education

Siqi Zhou (作者)

School of Mathematics, Guangdong University of Education

Jingying Guo (作者)

School of Mathematics, Guangdong University of Education

DOI：https://doi.org/10.63944/1gnmqw74

关键词：

Sedentary behavior, Sleep stage classification, Interpretable machine learning

已出版

2025-12-31

摘要: As wearable technologies and motion sensors advance, analyzing high-frequency accelerometer data has become a powerful tool for understanding physical activity, sleep patterns, and sedentary behavior. However, extracting reliable insights from such data remains challenging due to signal noise, individual variability, and the lack of labeled information. This study introduces an interpretable multi-task framework that integrates machine learning, unsupervised clustering, and temporal modeling to address these limitations. Using 24-hour wrist-worn accelerometer data from 100 adults, we applied a sliding-window approach and extracted time, frequency, and circadian rhythm features. For energy expenditure, Random Forest achieved the best prediction of metabolic equivalent (MET) values with R² = 0.967 and SMAPE = 9.54%, outperforming Elastic Net, XGBoost, and LightGBM. In sleep analysis, clustering via K-Means and GMM, refined by a Hidden Markov Model, successfully identified wakefulness, light sleep, and deep sleep without labeled ground truth. Sedentary risk was assessed using MET thresholds (<1.5) and Markov based state transitions, leading to a rule-based alert system for detecting prolonged inactivity. The proposed system combines accuracy with interpretability and offers strong potential for applications in mobile healthcare, eldercare, and chronic disease management.

参考文献

1. Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., Bassett, D. R., Tudor-Locke, C., ... & Leon, A. S. (2011). 2011 Compendium of Physical Activities: A second update of codes and MET values. Medicine & Science in Sports & Exercise, 43(8), 1575–1581. https:// doi.org/10.1249/MSS.0b013e31821ece12 DOI: https://doi.org/10.1249/MSS.0b013e31821ece12
2. AlEroud, A., & Karabatis, G. (2019). Probabilistic modeling for human behavior analysis: Hidden Markov and conditional random fields. Expert Systems with Applications, 130, 25–37. https://doi.org/10.1016/j.eswa.2019.04.026 DOI: https://doi.org/10.1016/j.eswa.2019.04.026
3. Bai, Y., Welk, G. J., Nam, Y. H., & Lee, J. A. (2016). Comparison of wrist and hip accelerometers under free-living conditions. BMC Public Health, 16, 1022. https://doi.org/10.1186/s12889-016-3661-1 DOI: https://doi.org/10.1186/s12889-016-3661-1
4. Banos, O., Galvez, J. M., Damas, M., Pomares, H., & Rojas, I. (2014). Feature extraction from wearable sensor data for activity recognition: A systematic review. Sensors, 14(6), 10080–10110. https://doi.org/10.3390/s140610080 DOI: https://doi.org/10.3390/s140609995
5. Bao, L., & Intille, S. S. (2004). Activity recognition from user-annotated acceleration data. In Pervasive Computing (pp. 1–17). Springer. https://doi.org/10.1007/978-3-540-24646-6_1 DOI: https://doi.org/10.1007/978-3-540-24646-6_1
6. Chastin, S. F. M., et al. (2015). Interventions to reduce sedentary behaviour in 0–5-year-olds: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 12, 127. https://doi.org/10.1186/s12966-015-0272-7 DOI: https://doi.org/10.1186/s12966-015-0272-7
7. Cole, R. J., Kripke, D. F., Gruen, W., Mullaney, D. J., & Gillin, J. C. (1992). Automatic sleep/wake identification from wrist activity. Sleep, 15(5), 461–469. https://doi.org/10.1093/sleep/15.5.461 DOI: https://doi.org/10.1093/sleep/15.5.461
8. Consolvo, S., McDonald, D. W., & Landay, J. A. (2009). Theory-driven design strategies for technologies that support behavior change in everyday life. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 405–414. https://doi. org/10.1145/1518701.1518766 DOI: https://doi.org/10.1145/1518701.1518766
9. Diaz, K. M., et al. (2017). Patterns of sedentary behavior and mortality in US middle-aged and older adults: A national cohort study. Annals of Internal Medicine, 167(7), 465–475. https://doi.org/10.7326/M17-0212 DOI: https://doi.org/10.7326/M17-0212
10. Ellis, K., Kerr, J., Godbole, S., Staudenmayer, J., & Lanckriet, G. (2014). Hip and wrist accelerometer algorithms for free-living behavio classification. Medicine & Science in Sports & Exercise, 46(9), 2016–2023. https://doi.org/10.1249/MSS.0000000000000288 DOI: https://doi.org/10.1249/MSS.0000000000000288
11. Ghose, D., Mitra, R., & Das, S. (2021). Sleep/wake classification from accelerometer data using hidden Markov model. Sensors, 21(1), 1–15. https://doi.org/10.3390/s21010001 DOI: https://doi.org/10.3390/s21010001
12. Giggins, O. M., Persson, U. M., & Caulfield, B. (2013). Biofeedback in rehabilitation. Sensors, 13(11), 14677–14688. https://doi.org/10.3390/s131114677
13. Hammerla, N. Y., Halloran, S., & Plötz, T. (2016). Deep, convolutional, and recurrent models for human activity recognition using wearables. IJCAI Proceedings, 1533–1540. https://arxiv.org/abs/1604.08880
14. Long, X., Fonseca, P., Foussier, J., Haakma, R., & Aarts, R. M. (2014). Modeling sleep and wakefulness in a home setting using wrist actigraphy and heart rate. IEEE EMBC, 4974–4977. https://doi.org/10.1109/EMBC.2014.6944670 DOI: https://doi.org/10.1109/EMBC.2014.6944670
15. Molnar, C. (2022). Interpretable machine learning: A guide for making black box models explainable. https://christophm.github.io/interpretable-ml-book/
16. Montoye, A. H. K., Moore, R. W., Bowles, H. R., Korycinski, R., & Pfeiffer, K. A. (2016). Reporting accelerometer methods in physical activity intervention studies. Journal of Physical Activity and Health, 13(6), 726–733. https://doi.org/10.1123/jpah.2015-0286
17. Owen, N., Healy, G. N., Matthews, C. E., & Dunstan, D. W. (2010). Too much sitting: The population health science of sedentary behavior. Mayo Clinic Proceedings, 85(12), 1138–1141. https://doi.org/10.4065/mcp.2010.0444 DOI: https://doi.org/10.4065/mcp.2010.0444
18. Patel, S., Park, H., Bonato, P., Chan, L., & Rodgers, M. (2012). A review of wearable sensors and systems with application in rehabilitation. Journal of NeuroEngineering and Rehabilitation, 9, 21. https://doi.org/10.1186/1743-0003-9-21 DOI: https://doi.org/10.1186/1743-0003-9-21
19. Piwek, L., Ellis, D. A., Andrews, S., & Joinson, A. (2016). The rise of consumer health wearables: Promises and barriers. PLOS Medicine, 13(2), e1001953. https://doi.org/10.1371/journal.pmed.1001953 DOI: https://doi.org/10.1371/journal.pmed.1001953
20. Rabiner, L. R. (1989). A tutorial on hidden Markov models and selected applications in speech recognition. Proceedings of the IEEE, 77(2), 257–286. https://doi.org/10.1109/5.18626 DOI: https://doi.org/10.1109/5.18626
21. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why should I trust you?”: Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD, 1135–1144. https://doi.org/10.1145/2939672.2939778 DOI: https://doi.org/10.1145/2939672.2939778
22. Sadeh, A. (2011). The role and validity of actigraphy in sleep medicine: An update. Sleep Medicine Reviews, 15(4), 259–267. https://doi.org/10.1016/j.smrv.2010.10.001 DOI: https://doi.org/10.1016/j.smrv.2010.10.001
23. Staudenmayer, J., Pober, D., Crouter, S., Bassett, D., & Freedson, P. (2009). An artificial neural network to estimate physical activity energy expenditure. Journal of Applied Physiology, 107(4), 1300–1307. https://doi.org/10.1152/japplphysiol.00465.2009 DOI: https://doi.org/10.1152/japplphysiol.00465.2009
24. Tremblay, M. S., Carson, V., Chaput, J. P., et al. (2017). Sedentary Behavior Research Network (SBRN)–Terminology consensus project. International Journal of Behavioral Nutrition and Physical Activity, 14(1), 75. https://doi.org/10.1186/s12966-017-0525-8 DOI: https://doi.org/10.1186/s12966-017-0525-8
25. Walch, O. J., Huang, Y., Forger, D. B., & Goldstein, C. A. (2019). Sleep stage prediction with raw acceleration and photoplethysmography using deep learning. npj Digital Medicine, 2, 101. https://doi.org/10.1038/s41746-019-0174-8
26. Freedson, P. S., Melanson, E., & Sirard, J. (1998). Calibration of the Computer Science and Applications, Inc. accelerometer. Medicine and Science in Sports and Exercise, 30(5), 777–781. https://doi.org/10.1097/00005768-199805000-00021 DOI: https://doi.org/10.1097/00005768-199805000-00021
27. Rosenberger, M. E., Buman, M. P., Haskell, W. L., McConnell, M. V., & Carstensen, L. L. (2016). Twenty-four hours of sleep, sedentary behavior, and physical activity with nine wearable devices. Medicine and Science in Sports and Exercise, 48(3), 457–465. https://doi. org/10.1249/MSS.0000000000000778 DOI: https://doi.org/10.1249/MSS.0000000000000778
28. Zhang, Z., Pi, Z., & Liu, B. (2015). TROIKA: A general framework for heart rate monitoring using wrist-type photoplethysmographic signals during intensive physical exercise. IEEE Transactions on Biomedical Engineering, 62(2), 522–531. https://doi.org/10.1109/ TBME.2014.2359372 DOI: https://doi.org/10.1109/TBME.2014.2359372
29. Migueles, J. H., Cadenas-Sanchez, C., Ekelund, U., Nyström, C. D., Mora-Gonzalez, J., Löf, M., & Ortega, F. B. (2017). Accelerometer data collection and processing criteria to assess
physical activity and other outcomes: A systematic review. Sports Medicine, 47(9), 1821–1845. https://doi.org/10.1007/s40279-017-0716-0 DOI: https://doi.org/10.1007/s40279-017-0716-0
30. Troiano, R. P., Berrigan, D., Dodd, K. W., Masse, L. C., Tilert, T., & McDowell, M. (2008). Physical activity in the United States measured by accelerometer. Medicine and Science in Sports and Exercise, 40(1), 181–188. https://doi.org/10.1249/mss.0b013e31815a51b3 DOI: https://doi.org/10.1249/mss.0b013e31815a51b3

Interpretable Modeling of Physical Activity, Sleep, and Sedentary Behavior from High-Frequency Wearable Data

版权所有 (c) 2025 國際人文社科研究

如何引用