Model-based Approaches to Information Diffusion Analysis in Social Networks
Nowadays the massive amount of information available from social networking sites makes it possible to build models for analysing the underlying social networks of interest.
Among the many aspects of online social network analysis, the study of information diffusion, that is, how information is propagated throughout the social networks, is a fundamental step to better understand the mechanisms of the social systems. Models built for information diffusion can be applied to a number of fields, such as recommendation systems and making marketing strategies.
In this project we will explore the use of model-based approaches to study the dynamics of information diffusion in social networks. We aim to automatically infer qualitative and semi-quantitative models of dynamic models from incomplete knowledge and imperfect data. Such models will have both predictive and explanatory power.
In this project we want to answer the following questions:
1. How to deal with large-scale model spaces and develop efficient model searching algorithms.
2. How to perform model selection given a set of candidate models.
3. Whether can we learn ab initio models from existing knowledge and social network data or learn models through composition of model fragments?
4. How can the learnt information diffusion models contribute to a better understanding of the underlying social system?
To achieve this we will first collect and analyse the social network data through publicly available packages and APIs. We will then use immune-inspired model learning algorithms to perform effective search on model spaces. We will also build a model library composed of model fragments for model composition. Finally we will assess the learnt models by applying Bayesian model selection techniques.
The successful applicant should have, or expect to have, an Honours Degree at 2.1 or above (or equivalent) in Computing Science or related disciplines. Knowledge - Essential: machine learning basics; social network analysis basics; programming in Java, Python, Ruby, or R. Desirable: modelling learning and simulation; Bayesian model selection; non-linear dynamics
The other supervisor on this project is Dr C Lin, Computing Science.
There is no funding attached to this project, it is for self-funded students only.
Formal applications can be completed online: http://www.abdn.ac.uk/postgraduate/apply. You should apply for PhD in Computing Science, to ensure that your application is passed to the correct College for processing. Please ensure that you quote the project title and supervisor on the application form.
Informal inquiries can be made to Dr W Pang (email@example.com) with a copy of your curriculum vitae and cover letter. All general enquiries should be directed to the Graduate School Admissions Unit (firstname.lastname@example.org).