Distributed Community Detection in Mobile Social Network

Abstract

The mobile social network is a type of delay tolerant network of mobile devices in which there is no end-to-end path available in advance for communication. It works on the principle of a store-carry-forward mechanism. The community is a very useful property of mobile social network as the human is a social animal and they like to live in a community. Using such social characteristics for detecting and using their community structure, we can enable them to communicate in an effcient manner. We have analysed various community detection methods and found the methods suitable for social networks. Such methods could be useful while creating our own algorithms for detecting communities or groups in mobile social networks. We have also analysed various existing distributed community detection algorithm in mobile social networks based on important parameters like complexity and type of community detected. Such analyses will help in discovering the strength and shortcoming of various mod- ern algorithms, which can be improved while preparing my algorithm. As the mobile social network is a self-organizing real network working on high resource constraint mobile devices, it is necessary to enable each mobile device to detect its own commu- nity with minimal information, computation and space requirements. This is a very challenging task and very little work is done in it. So there is an immense opportunity available for research in this area. We have implemented an existing algorithm which is an improvement of a popular algorithm. We have simulated it on a realistic like mobility model and found that it is not giving better results. So We have proposed and developed an improved algorithm, implemented and simulated it on the same realistic like mobility model. The results of the proposed algorithm prove that it is effective than the existing algorithm, as it provides better results.