User-Based Collaborative Filtering Mobile Health System
summary
A user-based collaborative filtering mobile health system uses a ten-week TILES study of 212 hospital workers with wearables, an audio phone sensor, OMsignal clothing, Fitbit, Owl-in-One and Minew to form cohorts by maximizing similarity across health labels, variable relationships, and sensor data. New users give five health labels that are discarded after cohort assignment. Cohort-based models enable cold-start personalized prediction, outperform baselines and beat individual models trained on the same five labels, showing usable, interpretable, privacy-aware results.
AI-written summary · about
related talks
39mBridging health inequalities through machine learning
14mDoes fitness data make the average person healthier?
19mWhat Healthcare Can Learn from Netflix
09mAn Activity Recognition System for Taking Medicine Using In-The-Wild Data to Promote Medication ...
01h 04mTake Control of the Data of You
22mReimaging Healthcare
26mMachine Learning for Personalised Health
27mBuild Your Own ML Data Feedback Loop