Deep quantile and deep composite triplet regression

DOI10.1016/j.insmatheco.2023.01.001OpenAlexW4317951323MaRDI QIDQ2685516

Michael Merz, Mario V. Wüthrich, Tobias Fissler

Publication date: 22 February 2023

Published in: Insurance Mathematics \& Economics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2112.03075

zbMATH Keywords

Bregman divergenceelicitabilityconditional tail expectationproper scoring ruleneural network regressionconsistent loss functionquantile and expected shortfall regressionsplicing model

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08) Applications of statistics to actuarial sciences and financial mathematics (62P05) Actuarial mathematics (91G05)

Related Items (1)

Conditionally Elicitable Dynamic Risk Measures for Deep Reinforcement Learning

Uses Software

• ElemStatLearn
• QRM
• Keras

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• A joint quantile and expected shortfall regression framework
• Making and Evaluating Point Forecasts
• Higher order elicitability and Osband's principle
• Eliciting production possibilities from a well-informed manager
• Seven proofs for the subadditivity of expected shortfall
• Elicitability and backtesting: perspectives for banking regulation
• Joint generalized quantile and conditional tail expectation regression for insurance risk analysis
• On the elicitability of range value at risk
• Severity modeling of extreme insurance claims for tariffication
• Order-sensitivity and equivariance of scoring functions
• Gamma mixture density networks and their application to modelling insurance claim amounts
• Autocalibration and Tweedie-dominance for insurance pricing with machine learning
• Composite Lognormal–Pareto model with random threshold
• Simultaneous multiple non-crossing quantile regression estimation using kernel constraints
• Quantile and Probability Curves Without Crossing
• Regression Quantiles
• Asymptotic Statistics
• Fat-Tailed Regression Modeling with Spliced Distributions
• Elicitation complexity of statistical properties
• A GENERALISED PROPERTY EXPOSURE RATING FRAMEWORK THAT INCORPORATES SCALE-INDEPENDENT LOSSES AND MAXIMUM POSSIBLE LOSS UNCERTAINTY
• A New Class of Severity Regression Models with an Application to IBNR Prediction
• Extending composite loss models using a general framework of advanced computational tools
• Strictly Proper Scoring Rules, Prediction, and Estimation
• Modeling actuarial data with a composite lognormal-Pareto model
• On composite lognormal-Pareto models
• DISTRIBUTION‐INVARIANT RISK MEASURES, INFORMATION, AND DYNAMIC CONSISTENCY
• Elicitation of Personal Probabilities and Expectations
• Random forests

This page was built for publication: Deep quantile and deep composite triplet regression