Passive Health Monitoring Using Large Scale Mobility Data
summary
Uses mobility traces from 2112 users collected for about two months, extracts 16 features (radius of gyration, displacement std, visit entropy, POI counts), and trains an XGBoost classifier to predict hospital visits achieving AUC=0.79. Interprets model with SHAP values to show lower mobility, lower visit diversity, fewer sports facilities and higher entertainment/fast-food density around visited places increase hospital-visit likelihood, with age and income modulating which features matter most.
AI-written summary · about
related talks
19mDigital Trace Data to Measure Human Mobility
17mKnow Thyself: Introspective Personal Data Mining
01h 19mOpportunities and risks of big data in disease surveillance.
07mBig Data and Cardiac Rehabilitation
10mIntelligence in the Age of Plethora
29mUser Motivation: Refining Customer Segments with Location
45mGoogle Health: Connecting Mobile Patients
24mBuilding Aggregated Solutions to Scale Your Geospatial Strategies