Predicting Successful Memes using Network and Community Structure [on Twitter]

Lilian Weng, Filippo Menczer, Yong-Yeol Ahn from Cornell University have created a model that can take a small amount of tweets and tell 2 months in advance whether the tweets will go viral and become a meme or not.

This is a network based model, that takes into account:

connectivity: number of early adopters, size of first and second surfaces (uninfected neighbours of early adopters);
distance: path length between consecutive users, variability in path length and maximum path length between any 2 adopters;
community features: number of communities with at least 1 adopter, how tweets or adopters of a given meme are scattered or concentrated across communities and intra-community interaction;
growth rate features: time between steps in the path and the variability of this time.

Their model is compared to 5 other models and comes out favoribly.

Whether the model can be adapted to other social networks is unclear.

[1403.6199] Predicting Successful Memes using Network and Community Structure.

Organisational Structures | Technology and Science | Military, IT and Lifestyle consultancy | Social, Broadcast & Cross Media | Flying aircraft

Leave a Reply