Querying cyber-networks using words

Abstract

Cyber-networks are characterized by two distinct types of nodes—devices and users—and by voluminous interactions in real time. All computational intelligence approaches to the analysis of these networks must deal with complexities of these interactions and their high data rates. This paper describes a computationally feasible approach to querying large cyber-networks using word-based queries. The attribute memberships and connection strengths in these cyber-networks are described granularly using appropriate vocabularies of words, where the words themselves are modeled using interval type-2 (IT2) fuzzy membership functions (MF) on an appropriate scale. By employing precomputation and storage of these word representations and queries, automated monitoring functions in large cyber-networks can be performed in real time via simple arithmetic calculations. We provide an illustrative example using data from a real cyber-network.