Analyzing the characteristics of a leader in a diffusion network

In today’s social and economic networks understanding the role that a single node has in influencing others plays an important role for policy-makers. In particular, we are interested in understanding what characteristics of the network make an initial injection point (the first persons to spread the information) impactful in the network. In this project, you will be hands-on in data analysis for analyzing information diffusion and will use recent models from network science to analyze the impact of the initial seeding nodes in influencing their peers. With the aim to get a better understanding of the network dynamics, you will be using tools from epidemic diffusion and graph signal processing. Understanding such dynamics and the role of the leaders within a network is highly preferred by advertisement-based companies that operate on online social networks. Moreover, you will gain expertise in tools including network science, epidemic diffusion, and graph signal processing. This knowledge will enrich your background gained during the master classes and prepare you for the job market.

Assignment

There are already a few relevant works on analyzing information diffusion on networks and the impact of the individual in driving such diffusion processes. In this project, your task will be to analyze a few models from the perspective of a single node to find out what characteristics increase its influence in the network. You will work both with mathematical models and real data. Therefore, this project is appropriate for a student that enjoys data analysis and data mining type of research. In addition, you will use recent tools from graph signal processing to understand the correlation between the type of the signal being diffused and the network topology. More in particular, you will adopt the following approach: (a) Perform a literature survey on information diffusion on networks. (b) Use relevant diffusion models to characterize the impact of the single node in the diffusion process. (c) Leverage recent advances in graph signal processing to find relations between the type of diffused signal and the network topology. (e) Evaluate the findings with real data and compare it with other state-of-the-art results.

Requirements

To be able to successfully complete such task, you should fulfill the following requirements: - be a self-motivated student interested to be hands on in data analysis and data mining - have a good linear algebra and mathematical background; - have knowledge of network analysis, statistical signal processing, and optimization theory - advanced coding skills in Matlab, Python (preferred), R, on any other programming language.

Contact

prof.dr.ir. Geert Leus

Signal Processing Systems Group

Department of Microelectronics

Last modified: 2019-02-08