University of Twente Student Theses


Influence Maximization in Social Networks by Injecting Memes

Sluiter, N.I. (2019) Influence Maximization in Social Networks by Injecting Memes.

[img] PDF
Abstract:The increasing importance of social networks opens up interesting discussions concerning the networks. This matter is extremely interesting in the marketing area, which influencer should companies approach for launching a new product, especially when there is limited budget? This poses several questions, which person in a social network should be chosen for initial injection of memes or a new product? To what extent are the node features crucial in predicting the most influential node? Can we, based on the results of the predicted node with the maximum influence, machine-learn a model to predict the most influential nodein the future? Data sets for the influence and values fornode centralities are calculated using the independent cas cade model and the networkx library respectively. Both linear and non-linear regressors are trained and tested using the data sets, resulting in R2 scores that define the accuracy of the regressor. Knowing which features maximize the R2 scores is useful for accurately predicting the influence in the future using the machine-learned model.Current Flow Closeness is the dominant feature for maximizing the accuracy as a single feature and in combinations of multiple features. Degree and Closeness give the worst accuracy.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page