ADAPTIVE REGULARIZED BAYESIAN HIERARCHICAL MODELLING FOR COMPLEX MULTIVARIATE DATA ANALYSIS
DOI:
https://doi.org/10.69980/2n38my80Keywords:
Bayesian hierarchical modelling, adaptive regularization, multivariate analysis, human activity recognition, high-dimensional dataAbstract
Conventional classification methods are faced with challenges of the high-dimensional multivariate datasets, strong correlation between predictors, and subject-level heterogeneity. In this study, an adaptive regularized Bayesian hierarchical modeling framework for complex multivariate data analysis using smartphone-based human activity data are developed and evaluated. Data comprised 10299 data points, 561 features extracted from the sensor data, 30 subjects, and 6 classes of activities, and a predefined split of the data was 7352 training and 2947 test data, which was subject independent. The structure was examined using multivariate structure analysis with principal component analysis, and those that predict the best were chosen through training-only feature screening. An adaptive coefficient shrinkage was added to the proposed model, and the model was compared to non-adaptive hierarchical Bayesian, adaptive fixed-effects Bayesian and multinomial L2 logistic models. On the remaining subjects, the proposed scheme attained 89.99% accuracy, 0.899 balanced accuracy, 0.905 macro precision, 0.899 macro recall and 0.900 macro F1 score. It also achieved a log loss of 0.241 and a multi-class Brier score of 0.144, which are better than the performance of the comparison models reported. The stability of the results from the predictive estimates was confirmed by subject-cluster bootstrap analysis. The greatest residual errors were between sitting and standing positions, and laying was considered a perfect classification. The results suggest that adaptive regularization along with hierarchical Bayesian structure can be efficiently used to handle predictor redundancy and clustered observations with a good out-of-sample classification performance in complex multivariate settings.
References
1. Ahmed, N., Rafiq, J. I., & Islam, M. R. (2020). Enhanced human activity recognition based on smartphone sensor data using hybrid feature selection model. Sensors, 20(1), 317.
2. Bhadra, A., Datta, J., Li, Y., & Polson, N. (2020). Horseshoe regularisation for machine learning in complex and deep models 1. International Statistical Review, 88(2), 302-320.
3. Brery, A. F., Gallardo-Antolín, A., Fakhry, M., & González-Carrasco, I. (2026). Deep learning for human activity recognition: A comprehensive review of architectures, performance, and challenges across five sensory datasets. Expert Systems with Applications, 323, 132415.
4. Bürkner, P. C., Gabry, J., & Vehtari, A. (2021). Efficient leave-one-out cross-validation for Bayesian non-factorized normal and Student-t models: P.-C. Bürkner et al. Computational Statistics, 36(2), 1243-1261.
5. Canale, A., Schiavon, L., & Stolf, F. (2026). Identifiable sparse Bayesian factorizations via meta regression. Statistics & Probability Letters, 110772.
6. Dentamaro, V., Gattulli, V., Impedovo, D., & Manca, F. (2024). Human activity recognition with smartphone-integrated sensors: A survey. Expert Systems with Applications, 246, 123143.
7. Donegan, C., Chun, Y., & Hughes, A. E. (2020). Bayesian estimation of spatial filters with Moran’s eigenvectors and hierarchical shrinkage priors. Spatial Statistics, 38, 100450.
8. Ferrari, A., Micucci, D., Mobilio, M., & Napoletano, P. (2021). Trends in human activity recognition using smartphones. Journal of Reliable Intelligent Environments, 7(3), 189-213.
9. Garcia-Gonzalez, D., Rivero, D., Fernandez-Blanco, E., & Luaces, M. R. (2023a). Deep learning models for real-life human activity recognition from smartphone sensor data. Internet of Things, 24, 100925.
10. Garcia-Gonzalez, D., Rivero, D., Fernandez-Blanco, E., & Luaces, M. R. (2023b). New machine learning approaches for real-life human activity recognition using smartphone sensor-based data. Knowledge-Based Systems, 262, 110260.
11. Greenacre, M., Groenen, P. J., Hastie, T., d’Enza, A. I., Markos, A., & Tuzhilina, E. (2022). Principal component analysis. Nature Reviews Methods Primers, 2(1), 100.
12. Griffin, J. E., & Brown, P. J. (2021). Bayesian global-local shrinkage methods for regularisation in the high dimension linear model. Chemometrics and Intelligent Laboratory Systems, 210, 104255.
13. Grinsztajn, L., Semenova, E., Margossian, C. C., & Riou, J. (2021). Bayesian workflow for disease transmission modeling in Stan. Statistics in medicine, 40(27), 6209-6234.
14. Kundu, S., Mallik, M., Saha, J., & Chowdhury, C. (2025). Smartphone based human activity recognition irrespective of usage behavior using deep learning technique. International Journal of Information Technology, 17(1), 69-85.
15. Lalwani, P., & Ganeshan, R. (2025). HAR: A stacked ensemble learning approach based on smartphone sensors for activity detection. Internet of Things, 30, 101487.
16. Qureshi, T. S., Shahid, M. H., Farhan, A. A., & Alamri, S. (2025). A systematic literature review on human activity recognition using smart devices: advances, challenges, and future directions. Artificial Intelligence Review, 58(9), 276.
17. Ramanujam, E., Perumal, T., & Padmavathi, S. J. I. S. J. (2021). Human activity recognition with smartphone and wearable sensors using deep learning techniques: A review. IEEE Sensors Journal, 21(12), 13029-13040.
18. Reshmi, S., & Ramanujam, E. (2024). An ensemble maximal feature subset selection for smartphone based human activity recognition. Journal of Network and Computer Applications, 226, 103875.
19. Reyes-Ortiz, J., Anguita, D., Ghio, A., Oneto, L., & Parra, X. (2013). Human activity recognition using smartphones [Data set]. UCI Machine Learning Repository. https://doi.org/10.24432/C54S4K
20. Reyes-Ortiz, J., Anguita, D., Ghio, A., Oneto, L., & Parra, X. (2013). Human activity recognition using smartphones [Data set]. UCI Machine Learning Repository. https://doi.org/10.24432/C54S4K
21. Straczkiewicz, M., James, P., & Onnela, J. P. (2021). A systematic review of smartphone-based human activity recognition methods for health research. NPJ Digital Medicine, 4(1), 148.
22. Thakur, D., & Biswas, S. (2022). An integration of feature extraction and guided regularized random forest feature selection for smartphone based human activity recognition. Journal of Network and Computer Applications, 204, 103417.
23. Thakur, D., & Biswas, S. (2023). Guided regularized random forest feature selection for smartphone based human activity recognition: D. Thakur, S. Biswas. Journal of Ambient Intelligence and Humanized Computing, 14(7), 9767-9779.
24. Tian, Y., & Zhang, Y. (2022). A comprehensive survey on regularization strategies in machine learning. Information Fusion, 80, 146-166.
25. Uddin, M. Z., & Soylu, A. (2021). Human activity recognition using wearable sensors, discriminant analysis, and long short-term memory-based neural structured learning. Scientific Reports, 11(1), 16455.
26. Veenman, M., Stefan, A. M., & Haaf, J. M. (2024). Bayesian hierarchical modeling: An introduction and reassessment. Behavior Research Methods, 56(5), 4600-4631.
27. Vehtari, A., Gelman, A., Simpson, D., Carpenter, B., & Bürkner, P. C. (2021). Rank-normalization, folding, and localization: An improved R ̂ for assessing convergence of MCMC (with discussion). Bayesian analysis, 16(2), 667-718.
28. Wei, R., & Ghosal, S. (2020). Contraction properties of shrinkage priors in logistic regression. Journal of Statistical Planning and Inference, 207, 215-229.
29. Zhang, S., Li, Y., Zhang, S., Shahabi, F., Xia, S., Deng, Y., & Alshurafa, N. (2022a). Deep learning in human activity recognition with wearable sensors: A review on advances. Sensors, 22(4), 1476.
30. Zhang, Y. D., Naughton, B. P., Bondell, H. D., & Reich, B. J. (2022b). Bayesian regression using a prior on the model fit: The R2-D2 shrinkage prior. Journal of the American Statistical Association, 117(538), 862-874.
