Prediction of Elderly Depression Based on an Improved Conditional Mixture of Experts Model

DOI

10.2991/978-94-6239-721-7_11How to use a DOI?

Keywords

Elderly depression; Ensemble learning; Conditional Mixture of Experts; CFPS

Abstract

Elderly depression, often underdiagnosed due to subtle symptoms, diminishes quality of life. This study develops a detection model using ensemble learning on multidimensional features of older adults from the China Family Panel Studies (CFPS). The dataset includes over 1,200 variables, with depression defined as CESD-8 score > 16. SMOTE addressed class imbalance, and a hybrid Boruta-Lasso strategy was used for feature selection. An improved Conditional Mixture of Experts (MoE) model was proposed. On the test set, the Improved MoE achieved an F1 score of 0.578 and recall of 74.36%; on the original (non-SMOTE) set, it maintained a recall of 68.2% and F1 of 0.521, outperforming all baselines with robustness confirmed by stratified cross-validation (95% CI). Health status, subjective well-being, and meaning in life were identified as key predictors. This study provides an efficient, interpretable tool for community-based elderly depression screening.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Lin Tan
PY  - 2026
DA  - 2026/07/06
TI  - Prediction of Elderly Depression Based on an Improved Conditional Mixture of Experts Model
BT  - Proceedings of the 2026 6th International Conference on Public Management and Intelligent Society (PMIS 2026)
PB  - Atlantis Press
SP  - 110
EP  - 119
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-721-7_11
DO  - 10.2991/978-94-6239-721-7_11
ID  - Tan2026
ER  -