A. Saavedra-Nieves, E. Algaba
Risk analysis and cooperative game theory are strongly connected, see Cox (2009). Through the use of well-known solution concepts for TU-games, new risk measures can be proposed in networks. The position value (Borm et al., 1992) gives a measure of the relevance of a player in a network, integrating additionally the degree measure of each player in it. In real world, it is important to consider the natural influence of connections of a player in a network. However, until now, its applications were very limited due to the high computational complexity. In this talk, we provide a sampling-based method to estimate the position value, which is analysed in terms of the theoretical properties of the resulting estimator. Besides, we establish statistical results for bounding the absolute error in this approximation. To illustrate the variety of problems that can be analyzed in this framework, we applied it in two different settings, the Spanish national team and the Zerkani network.
Palabras clave: Position value, communication situation, sampling, covert networks, ranking.
Programado
GT19.GAMES4. Game Theory Working Group: Session III in honor of Prof. Stef Tijs
10 de noviembre de 2023 09:30
CC2: Sala Conferencias