Intelligent system for dynamically providing suggestions for meetup based on proximity

This paper presents a method for smart meetup based on the users´ proximity and using contextual information present in a mobile device and information publicly available on the Internet. The system uses the user´s data from the device and the user´s historical data available online to accurately suggest persons to meet up in close proximity to the user and also suggest activities to be done with them. Machine learning techniques and tools are used to derive the score for desirability of meetup. The input data consists of a variety of factors, including the user´s past behavior online, data on user´s profile in the device (call and usage logs), input from the device sensors and data about user´s contacts in proximity. While processing the input data, the system learns rules about how likely it is for the user to wish to meet the contact person. If the desirability for meet up is positive, the system gives suggestions for activities to do together, based on available data about businesses or activities in the area. If the desirability for meetup is negative, the system gives a warning to the user to avoid the contact in proximity. In this paper we provide server based and device based implementations for the system and analyze issues with its functioning.