| 摘要 |
Predicting a current destination comprises receiving user data and activity data. At least one non-routine event is determined from the user data, each non-routine event being defined by a respective event time and a respective predicted event location derived from a non-specific location reference (such as a colloquial or user specific location name, or an ambiguous location). At least one routine event is determined from user activity, each routine event being defined by an event time and an event location. Each event time is compared with the current time to determine a current event, and the predicted destination is determined using the event location or predicted event location corresponding to the current event. The user data may be a personal calendar comprising appointments or events. Information relevant to the predicted destination or a route to the predicted destination may be presented to a user. Non-specific location references may be correlated with location data included in the user data to train a machine learning algorithm to output predicted location data from the non-specific location references. The system may be implemented in a vehicle. |
