SOCIAL NETWORK IDENTIFICATION AND ANALYSIS USING CALL DETAIL RECORDS

Abstract

This thesis proposes comprehensive solution analysis of call detail records for discovering and analyzing social networks and prominence of an actor in the network. It is based on all the existing techniques and methodologies for telecom billing solutions. On the basis of existing systems database has been developed in SQL server 2005. The front end has been developed in Visual Studios 2005 in Visual C# environment. All the basic functionalities have been provided for billing systems. The basic functionalities include searching through CDR and generating required reports from the database. Also revenue generating customers have been identified. Finally methodologies have been proposed for identification and analysis of social networks from CDR. These methodologies include grouping the related mobile users using CDR’s by selecting a central node. Also networks analysis has been made on the basis of degree centrality and prestige to find the prominence of a user in the group or whole network or in cluster. Algorithms have been proposed for clustering and for network analysis. Finally simulation for the working of the grouping is developed. This simulation is quite user friendly. It shows the link information between actors in a group or cluster in quite a comprehensive way. It explains most of the functionality with sociogram generation.