Digital Phenotyping
The article says a prior study by the same team achieved an AUC of 0.77 for one-week suicidal risk prediction using passive digital data. The primary digital phenotyping indicators are geolocation and time-stamped Google Maps Timeline data…
1 sources - 5 claims
The article says a prior study by the same team achieved an AUC of 0.77 for one-week suicidal risk prediction using passive digital data. The primary digital phenotyping indicators are geolocation and time-stamped Google Maps Timeline data. Additional Google-derived features are used as proxies for activity, sleep-wake patterns, and social interaction. Google Takeout data are pseudonymised before secure storage on servers managed by Universidad Carlos III. SMARTomicS asks participants to export personal data through Google Takeout.