Speaker: Fernanda Gabrielli
Abstract: This talk explores how predicting missing links can change the results of game-theoretic centrality measures in networks. Using the Zerkani network as an example, we apply several link prediction methods and observe how they alter the β-measure and the Owen value, even under different community structures. We also quantify how robust these measures are to added links. The goal is to show how incomplete data—and the way we complete it—can significantly affect who appears most “important” in a network.
Fecha y hora: miércoles 19 de noviembre, 12:00hrs.
Lugar: Sala Seminario Felipe Álvarez (5to piso)
Link zoom: https://uchile.zoom.us/j/93417604828?pwd=eSHafQqVcAegttXqKH3swXZsMseRrn.1