Local Information Diffusion Patterns in Social and Traditional Media: The Estonian Case Study
Name
Oliver Soop
Abstract
Information has become more highly valued among companies and individuals than ever before. With this, the interest in how information diffuses among the entities in various structured networks has increased. A number of studies have been published on the diffusion process in real-life networks, such as web service network, citation networks, blog networks etc. Majority of researches have focused on one type of network - such as Facebook posts, Twitter tweets, Blogspot blog entries etc. A disadvantage of analysing a network containing entities from a single source is that it does not consider the outside influence on the diffusion. Recently, some papers have started to incorporate different networks in their study and as such have been able to analyse the effect of outside influence on the diffusion process.
This thesis aims to shed further light into the topic of information diffusion in a real world network containing entities from different sources, this is achieved by detection of relevant local topological and temporal information diffusion patterns. For topological pattern analysis, frequent subgraph mining techniques are used. Temporal patterns are extracted using time series clustering. The dataset used in this thesis is collected from the Estonian setting of mainstream online news media with comments and articles and from social media channels Twitter and Facebook. From this dataset the relations between the entities were extracted and a network for analysis of diffusion patterns was constructed. Temporal patterns reveal the high pace of information diffusion while topological patterns expose the important role of news media articles and Facebook posts in the information diffusion processes. The results of the thesis are applicable in cyber defence, online marketing and campaign management plus information impact estimation, just to mention a few application areas.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Peep Küngas
Defence year
2014