An innovative approach to detect Spam over Internet Telephony by applying artificial intelligence methods

Abstract

The dissertation thesis deals with the need to provide a new approach to a protection from Spam over Internet Telephony. It is not an entirely trivial matter to ensure the security of the VoIP services and attacks on telecommunication solutions, built on VoIP technology, grow with an increasing number of active users in last two decades. Spam over Internet Telephony (SPIT) is one of the attacks going to have a significant impact on the user experience. Various techniques have been presented to detect SPIT calls so far, however, nobody has devoted his/her research to the application of modern artificial intelligence algorithms on SPIT detection. It is my motivation to propose an innovative detection mechanism which is the primary goal of my dissertation thesis.

Due to nonexistence of active open-source projects dedicated to VoIP attacks, which I could use for creation of the suitable dataset, I had to develop my own tool. No project dispose of sophisticated SPIT detection, and many scientific research papers work with synthetic data, which behave according to the defined models. By processing a wide range of data from various sources, using different statistical and data mining tools, I focused my research on the behavioral characteristics of individual users in the telecommunications chain. I was able to determine the decisive factors for the design and innovative concept of a classification system that is able to distinguish between a regular user and a SPIT attacker. Among other things, this dissertation brings a new approach to the emulation SIP relations based on Markov chains, as well as the design of a new open-source honeypot with the advanced features.