ACTIVITY-BASED PERSONAL PROFILE INFERENCE

摘要

The aggregation of facts from various sources about an individual may produce an individual profile that may inform personalized services. However, a compilation of facts may be supplemented by monitoring activities of the individual and formulating inferences regarding the individual's individual details, and the confidence of such inferences. Accordingly, a device may compare the detected activities with a behavioral rule set indicating correlations between activities and inferred individual details (e.g., frequently spent weekday evenings and morning departures from a residence imply that the residence is the individual's home; frequent bicycling to work, chosen over other available modes of transportation, implies that the individual is a bicycling enthusiast) to add inferred individual details to the individual profile. Continued monitoring may enable updating based on changes to the individual details. Multiple profiles may be synchronized while respecting the individual's privacy, obtaining the individual's consent to share information, and automatically resolving information conflicts.