Project №26-25AI Engineering

Menopause Symptom-Trajectory Research Model

A women's-health research initiative needed an interpretable predictive-modeling pilot to differentiate meaningful menopause symptom-burden trajectories using public longitudinal research data. Delivering a working, interpretable reference model with documented discrimination, calibration, and limitations.

Back to archive
Menopause Symptom-Trajectory Research Model
Brief

Build an interpretable predictive-modeling pilot to differentiate clinically meaningful menopause symptom-burden trajectories from longitudinal research data.

Approach

Confirming the university IRB pathway and public research data-use process, building a variable dictionary and missingness map, defining trajectory groups, comparing interpretable baselines against gradient-boosted models, and evaluating calibration, subgroup performance, and feature attribution.

Outcome

A working, interpretable reference model on held-out research data, with documented discrimination, calibration, subgroup performance, and a reproducibility package — research and predictive-modeling scope only, with no clinical validation or diagnosis.

IRB-reviewed
University research pathway
Interpretable model
Documented calibration + limits
Reproducible
Code, notebooks, model card