direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Philipp Lorenz, M.Sc.


Institut für Theoretische Physik
ER 240 (Ernst Ruska Gebäude)
Sekr. EW 7-1
Institut für Theoretische Physik
Technische Universität Berlin
Hardenbergstr. 36
D-10623 Berlin, Germany

Temporal communities of hashtags

Click for interactive version

With more and more temporal data available, I develop tools to reliably track topics throughout time. To do so I create temporal hashtag co-occurence networks, apply community detection algorithms and track their history with a memory based approach.

Modeling the rise and fall of online topics

Comparisson of the empirical distributions of relative gains and losses in popularity with the model.

I can model the resulting dynamics using a simple preferential-attachment and detachment model, which incorporates ranking dynamics as well. The model parameters and the required configuration for a given dataset is informational with respect to the sociological and psychological mechanisms that drive the dynamics of popularity in different contexts.

Threshold models with repost and recovery

(PNG, 648,3 KB)
Comparison between empirical observed spreading of a community of hashtags and a simulated outbreak, using a threshold model with accumulation (repost) effects

The nouvelle aspects of complex contagion of opinions and trends, arising from social media. Mechanisms like reposting or disliking as well as new roles, like content-creators and followers, make extensions of existing models (e.g. threshold-models) necessary.

Zusatzinformationen / Extras