Quantiles as optimal point forecasts

By Tilmann Gneiting. 2011

228 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/9a6561fc7304ef54fa94f9978cf198240a255d6c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

Citations in the corpus

Wind Energy: Forecasting Challenges for Its Operational Management (2013, methodology): “The estimation of model parameters is then performed based on the check function criterion of Koenker and Bassett (1978) or any general scoring rules for quantiles (Gneiting and Raftery, 2007), instead of quadratic and likelihood-based criteria. An example approach to point forecasting of wind power”
Making and Evaluating Point Forecasts (2009, background): “For a discussion of their equivalence and historical comments, see Gneiting (2011).”
Bayesian structured additive distributional regression with an application to regional income inequality in Germany (2015, background): “This decomposition is especially helpful in situations where the quantile score can be interpreted as an economically relevant loss function [Gneiting (2011b)].”
Very‐short‐term probabilistic forecasting of wind power with generalized logit–normal distributions (2012, background): “Following the point of Gneiting (2011) among others, one should extract the optimal point forecasts from predictive densities based on the target evaluation score.”
Bayesian Structured Additive Distributional Regression (2013, background): “In many applications, e.g. insurances [Klein et al., 2013a, Heller et al., 2006] or weather forecasts [Gneiting and Ranjan, 2011], distributions with point masses at zero are of great interest.”
Forecasting: theory and practice (2020, background): “• The hinge/tick/pinball loss is minimised by the appropriate quantile (Gneiting, 2011b).”
On the importance of the long-term seasonal component in day-ahead electricity price forecasting (2016, background): “…in terms of the robust weekly-weighted mean absolute error (WMAE; see Weron, 2014) for point forecasts, the pinball loss function for probabilistic forecasts (Gneiting, 2011; Hong et al., 2016), and the Diebold and Mariano (1995) test for significant differences in forecasting…”
Using Stacking to Average Bayesian Predictive Distributions (with Discussion) (2017, methodology): “Adapting the notation of Gneiting and Raftery (2007), we label Y as the random variable on the sample space (Ω,A) that can take values on (−∞,∞).”
Short‐Term Wind Speed Forecasting for Power System Operations (2012, background): “However, Gneiting (2011a) pointed out that “This can lead to grossly misguided inferences, unless the loss function and the forecasting task are carefully matched”.”
Probabilistic Forecasts of Wind Power Generation Accounting for Geographically Dispersed Information (2014, background): “More generally speaking, the benefits of probabilistic forecasts can be justified by the fact that for a large class of decision making problems the optimal solution is directly linked to a specific quantile rather than the expectation of the future outcome [14].”
Recent advances in electricity price forecasting: A review of probabilistic forecasting (2016, background): “Pinball loss [87, 88] Winkler (interval) score [89] Diebold-Mariano [90, 91] Model confidence set [92] Forecast encompassing [93] Continuous Ranked Probability Score (CRPS) [16, 94] Logarithmic score [95] Diebold-Mariano [90, 91] Model confidence set [92] Forecast encompassing [93]”
Comparing Possibly Misspecified Forecasts (2020, background): “(Here and throughout, we will focus on strict consistency of a loss function, which in this section requires strict convexity of φ; see Gneiting (2011a) for discussion of consistency vs. strict consistency.)”
On the Comparison of Interval Forecasts (2018, background): “Gneiting (2011a), for example, discusses a scoring rule for the mode, but he does not find a scoring rule that selects the mode as the global optimum.”