Sensitivity measures based on scoring functions
From MaRDI portal
Publication:6167385
Abstract: We propose a holistic framework for constructing sensitivity measures for any elicitable functional of a response variable. The sensitivity measures, termed score-based sensitivities, are constructed via scoring functions that are (strictly) consistent for . These score-based sensitivities quantify the relative improvement in predictive accuracy when available information, e.g., from explanatory variables, is used ideally. We establish intuitive and desirable properties of these sensitivities and discuss advantageous choices of scoring functions leading to scale-invariant sensitivities. Since elicitable functionals typically possess rich classes of (strictly) consistent scoring functions, we demonstrate how Murphy diagrams can provide a picture of all score-based sensitivity measures. We discuss the family of score-based sensitivities for the mean functional (of which the Sobol indices are a special case) and risk functionals such as Value-at-Risk, and the pair Value-at-Risk and Expected Shortfall. The sensitivity measures are illustrated using numerous examples, including the Ishigami--Homma test function. In a simulation study, estimation of score-based sensitivities for a non-linear insurance portfolio is performed using neural nets.
Recommendations
- scientific article; zbMATH DE number 1222775
- Efficient estimation of sensitivity indices
- Order-sensitivity and equivariance of scoring functions
- Sensitivity in Metric Scaling and Analysis of Distance
- Sensitivity analyses and measurements for group decisions using weighted scoring rules
- Sensitivity function: properties and applications
- The characterisation of scoring functions
- A new measure based on sensitivity and specificity
- The score function approach for sensitivity analysis of computer simulation models
- Global sensitivity measures from given data
Cites work
- Copula theory and probabilistic sensitivity analysis: is there a connection?
- Data-based ranking of realised volatility estimators
- DISTRIBUTION‐INVARIANT RISK MEASURES, INFORMATION, AND DYNAMIC CONSISTENCY
- Elicitability and backtesting: perspectives for banking regulation
- Estimation of quantile oriented sensitivity indices
- Forecast evaluation of quantiles, prediction intervals, and other set-valued functionals
- Global sensitivity analysis and Wasserstein spaces
- Global sensitivity analysis: The primer
- Higher order elicitability and Osband's principle
- Is the mode elicitable relative to unimodal distributions?
- Making and evaluating point forecasts
- Measurability of functionals and of ideal point forecasts
- New sensitivity analysis subordinated to a contrast
- Of quantiles and expectiles: consistent scoring functions, Choquet representations and forecast rankings. With discussion and authors' reply
- On a multivariate copula-based dependence measure and its estimation
- On measures of dependence
- On the elicitability of range value at risk
- On the Relative Importance of Input Factors in Mathematical Models
- Order-sensitivity and equivariance of scoring functions
- Probabilistic sensitivity measures as information value
- Quantitative risk management. Concepts, techniques and tools
- Reverse sensitivity testing: what does it take to break the model?
- Risk functionals with convex level sets
- Robustness and sensitivity analysis of risk measurement procedures
- Sensitivity analysis based on Cramér-von Mises distance
- Sensitivity analysis: a review of recent advances
- Strictly Proper Scoring Rules, Prediction, and Estimation
- The geometry of proper scoring rules
- The mode functional is not elicitable
- Using Proper Divergence Functions to Evaluate Climate Models
Cited in
(10)- Sensitivity in Metric Scaling and Analysis of Distance
- Conditionally Elicitable Dynamic Risk Measures for Deep Reinforcement Learning
- Regression diagnostics meets forecast evaluation: conditional calibration, reliability diagrams, and coefficient of determination
- Fifty years of decision analysis in operational research: a review
- Convexity and measures of statistical association
- Robust elicitable functionals
- Differential quantile-based sensitivity in discontinuous models
- Optimal transport divergences induced by scoring functions
- A cheat sheet for Bayesian prediction
- Kullback-Leibler barycenter of stochastic processes
This page was built for publication: Sensitivity measures based on scoring functions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6167385)
