- AUTHORS: Elena Cerrato & Alberto Ibarrondo
- LANGUAGE: Spanish
- PROGRAMMING LANGUAGE: R
Social networks are continuously increasing their importance on the life of people. They allow us to communicate with user from any nation and be informed of everything going on around the globe. But not only do they favor the users, social networks also offer brand new channels for companies to invest in advertisement campaigns, reaching further than any traditional media. These campaigns are known as Viral Marketing, the basis of this project.
In order to benefit from all the advantages the social networks may present, the companies interested must discern which individuals should initially convince to spread the product/service, thus achieving maximum reach.
The scope of this project will be searching social networks for the group of users who will expand the campaign the furthest possible. Solving this problem will require social network modeling as a graph and studying optimum selection techniques related to both structural centrality and dynamic diffusion in networks.
After reviewing all the possibilities, the Lineal Threshold Model is selected for its better adaptation to the viral marketing environment. Once chosen, the optimization function is defined, taking into account the total number of reached nodes and the time invested for such.
With both model and function set, the functions are coded and a series of simulations are designed and carried out for real and randomly generated graphs. The results enable the selection of the best function for searching of optimal groups in networks. The project concludes with several recommendations for its use in real life scenarios.