This study constructs a health monitoring model for the early screening of AD that integrates environmental data and behavioral characteristics. Based on 174 452-dimensional handwriting feature samples from the DARWIN dataset, the median was filled with missing values, and Pearson correlation was used to screen the top 20 highly correlated features (strong positive correlation in the MOX group). The performance differences of PCA, UMAP, and LDA dimensionality reduction techniques were compared. Through the classification of eight machine learning algorithms, it was found that the random forest had the best performance after PCA dimensionality reduction (MSE=4.53, R²=0.9965, which was 61% better than linear regression). Feature analysis shows that the MOX4 related to handwriting pressure/time is the core indicator, and the weight of environmental features is less than 3%. The research provides a lightweight screening solution centered on handwriting features, verifies the efficiency of PCA in extracting the linear structure of high-dimensional data, and offers a methodological reference for non-invasive monitoring in intelligent elderly care. In the future, it can integrate multimodal data to build an early warning model.

Shuijin Rong, Wei Guo, Hao Liu. Research on Elderly Health Monitoring Based on Multiple Machine Learning Algorithms. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 7: 47-54. https://doi.org/10.25236/AJCIS.2025.080706.

[1] Zhang, M., Shi, Y., Zhou, B., Huang, Z., Zhao, Z., Li, C., ... & Li, Y. (2023). Prevalence, awareness, treatment, and control of hypertension in China, 2004-18: findings from six rounds of a national survey. Bmj, 380.

[2] Wang, L., Di, X., Yang, L., & Dai, X. (2020, November). Differences in the potential accessibility of home-based healthcare services among different groups of older adults: a case from shaanxi province, china. In Healthcare (Vol. 8, No. 4, p. 452). MDPI.

[3] Zanetti, O., Solerte, S. B., & Cantoni, F. (2009). Life expectancy in Alzheimer's disease (AD). Archives of gerontology and geriatrics, 49, 237-243.

[4] Alberdi, A., Aztiria, A., & Basarab, A. (2016). On the early diagnosis of Alzheimer's Disease from multimodal signals: A survey. Artificial intelligence in medicine, 71, 1-29.

[5] Li, A., Xue, C., Wu, R., Wu, W., Zhao, J., & Qiang, Y. (2025). Unearthing subtle cognitive variations: a digital screening tool for detecting and monitoring mild cognitive impairment. International Journal of Human–Computer Interaction, 41(4), 2579-2599.

[6] Xue, H., Yang, Q., & Chen, S. (2009). SVM: Support vector machines. In The top ten algorithms in data mining (pp. 51-74). Chapman and Hall/CRC.

[7] Chicco, D., Warrens, M. J., & Jurman, G. (2021). The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. Peerj computer science, 7, e623.

[8] Adler, W., & Lausen, B. (2009). Bootstrap estimated true and false positive rates and ROC curve. Computational statistics & data analysis, 53(3), 718-729.

[9] Maffezzoni, D., Barbierato, E., & Gatti, A. (2025). Data-Driven Diagnostics for Pediatric Appendicitis: Machine Learning to Minimize Misdiagnoses and Unnecessary Surgeries. Future Internet, 17(4), 147.