SEIZ Matters: Modelling the spread of concepts on Twitter

Back in the good old days of plentiful Twitter API access, I investigated how best to model the spread of new concepts on Twitter. Specifically, I tested the sociology-inspired SEIZ model versus the simpler SI model.

In a nutshell, I found that SEIZ does indeed matter (it provides a better fit) but that it’s equally important how you use it (it provides rich interpretations of the parameters).

While the work was solo-authored I am greatly indebted to conversations with my professor Fabio Trecca.

Recommended citation: Rystrøm, J. H. (2020). SEIZ Matters: Modelling the spread of concepts on Twitter. Journal of Language Works - Sprogvidenskabeligt Studentertidsskrift, 5(1), 78–91. Retrieved from https://tidsskrift.dk/lwo/article/view/121220
Download Paper