Symmetry, mutual dependence, and the weighted Shapley values

被引:12
|
作者
Casajus, Andre [1 ,2 ]
机构
[1] HHL Leipzig Grad Sch Management, Jahnallee 59, D-04109 Leipzig, Germany
[2] Dr Hops Craft Beer Bar, Eichendorffstr 7, D-04277 Leipzig, Germany
关键词
TU game; Weighted Shapley values; Symmetry; Mutual dependence; Weak differential marginality; Superweak differential marginality; BALANCED CONTRIBUTIONS; GAMES;
D O I
10.1016/j.jet.2018.09.001
中图分类号
F [经济];
学科分类号
02 ;
摘要
We pinpoint the position of the (symmetric) Shapley value within the class of positively weighted Shapley value to their treatment of symmetric versus mutually dependent players. While symmetric players are equally productive, mutually dependent players are only jointly (hence, equally) productive. In particular, we provide a characterization of the whole class of positively weighted Shapley values that uses two standard properties, efficiency and the null player out property, and a new property called superweak differential marginality. Superweak differential marginality is a relaxation of weak differential marginality (Casajus and Yokote, 2017). It requires two players' payoff for two games to change in the same direction whenever only their joint productivity changes, i.e., their individual productivities stay the same. In contrast, weak differential marginality already requires this when their individual productivities change by the same amount. The Shapley value is the unique positively weighted Shapley value that satisfies weak differential marginality. (C) 2018 Elsevier Inc. All rights reserved.
引用
收藏
页码:105 / 123
页数:19
相关论文
共 50 条
  • [21] P-Shapley: Shapley Values on Probabilistic Classifiers
    Xia, Haocheng
    Li, Xiang
    Pang, Junyuan
    Liu, Jinfei
    Ren, Kui
    Xiong, Li
    PROCEEDINGS OF THE VLDB ENDOWMENT, 2024, 17 (07): : 1737 - 1750
  • [22] Computing Shapley Values in the Plane
    Sergio Cabello
    Timothy M. Chan
    Discrete & Computational Geometry, 2022, 67 : 843 - 881
  • [23] Explaining Preferences with Shapley Values
    Hu, Robert
    Chau, Siu Lun
    Huertas, Jaime Ferrando
    Sejdinovic, Dino
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [24] On the failings of Shapley values for explainability
    Huang, Xuanxiang
    Marques-Silva, Joao
    INTERNATIONAL JOURNAL OF APPROXIMATE REASONING, 2024, 171
  • [25] Computing Shapley Values in the Plane
    Cabello, Sergio
    Chan, Timothy M.
    DISCRETE & COMPUTATIONAL GEOMETRY, 2022, 67 (03) : 843 - 881
  • [26] Consistent restricted Shapley values
    Derks, J
    Peters, H
    MATHEMATICAL SOCIAL SCIENCES, 1997, 33 (01) : 75 - 91
  • [27] Parallel axiomatizations of weighted and multiweighted Shapley values, random order values, and the Harsanyi set (vol 71, pg 5664, 2020)
    Besner, Manfred
    SOCIAL CHOICE AND WELFARE, 2020, 55 (01) : 213 - 214
  • [28] Weighted Aumann-Shapley pricing
    McLean, RP
    Sharkey, WW
    INTERNATIONAL JOURNAL OF GAME THEORY, 1998, 27 (04) : 511 - 523
  • [29] Weighted Aumann-Shapley pricing
    Richard P. McLean
    William W. Sharkey
    International Journal of Game Theory, 1998, 27 : 511 - 523
  • [30] Manifold Restricted Interventional Shapley Values
    Taufiq, Muhammad Faaiz
    Bloebaum, Patrick
    Minorics, Lenon
    INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 206, 2023, 206