Future Location Prediction of a Mobile User Using Historic Visiting Patterns

摘要

The ability of modern smartphones to provide us with real time location-based data is one of its most important features. Being able to predict a person’s future location based on the real time location data would be the next step in utilizing this functionality. Using this functionality, combined with machine learning one’s probable destination can be predicted with a reasonable accuracy. People don’t always use map-based navigation for the places they visit every day, like their work place or school and there may be significant traffic on the regular route taken, however, if our device knows where we’re headed, it can warn us beforehand and help us reroute. This functionality can also be used by cops to determine the future location of a criminal fleeing a crime scene.These features and functionalities can be implemented through various machine learning algorithms which are compared to determine the most accurate one. The proposed system can predict a user’s future location using the current location and time, learning from the user’s previously visited locations.