Accession Number : AD1026864

Title :   Partial Information Community Detection in a Multilayer Network

Descriptive Note : Technical Report

Corporate Author : Naval Postgraduate School Monterey United States

Personal Author(s) : Warnke,Scott D

Full Text :

Report Date : 01 Jun 2016

Pagination or Media Count : 131

Abstract : Identifying communities in a dark network is a potentially difficult task. The nature of dark networks, and their characteristic of concealing connections within the network, makes community detection an enterprise based on operations and decisions with only partial information. We take this concept of operation with only partial information, and extend it to our work by identifying communities within a dark network using only a single layer from the full multilayer network. Additionally, the concept of identification of terrorist networks within civilian populations is one of ever-increasing importance in our world today. We create a large multilayer synthetic network, and embed a known terrorist network in the larger synthetic network. We construct our synthetic network in a manner to ensure that our terrorist network is not unique, in order to make discovery of the terrorist network difficult. In this portion of our work we are concerned with identifying the entire terrorist network, not just a community within the terrorist network. We use known discovery algorithms to discover the terrorist network, and compare the results to modified algorithms introduced in this thesis and their ability to discover the terrorist network as quickly as possible

Descriptors :   algorithms , internet , probability , social networking services , random walk , graph theory , electronic mail , mathematical models , NETWORK ARCHITECTURE , INFORMATION EXCHANGE , TERRORISM , GROUP RELATIONS , MATRICES (MATHEMATICS)

Distribution Statement : APPROVED FOR PUBLIC RELEASE