Measures of the Shapley index for learning lower complexity fuzzy integrals

摘要: The fuzzy integral (FI) is used frequently as a parametric nonlinear aggregation operator for data or information fusion. To date, numerous data-driven algorithms have been put forth to learn the FI tasks like signal and image processing, multi-criteria decision making, logistic regression minimization of sum squared error (SEE) criteria in decision-level However, existing work has focused on learning densities (worth just individual inputs underlying measure (FM)) relative an imputation method (algorithm that assigns values remainder FM) full FM learned single (e.g., SSE). Only handful approaches investigated how some (logistic SSE) conjunction with second criteria, namely model complexity. Including complexity important because it allows us solutions are less prone overfitting we can lower solution’s cost (financial, computational, etc.). Herein, review compare different indices (capacity) We show there no global best. Instead, applications goals (contexts) what drives which index appropriate. In addition, new based functions Shapley index. Synthetic real-world experiments demonstrate range behavior these