Forecast Combination and Model Averaging Using Predictive Measures

By Jana Eklund, Sune Karlsson. 2007

195 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/067320f325d692258523e78825031ff0ec41a0af

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

In Management Science and Operations Research

Citations in the corpus

Combining forecast densities from VARs with uncertain instabilities (2010, methodology): “…recursive logarithmic score weights, RW, has many similarities with an approximate predictive likelihood approach (see Raftery & Zheng (2003), and Eklund & Karlsson (2007)).4 Given our definition of density fit, the model densities are combined using Bayes’ rule with equal (prior) weight on each…”
Forecast Combination and Bayesian Model Averaging - A Prior Sensitivity Analysis (2012, methodology): “Eklund and Karlsson (2007) recommend to reserve 75% of the observations to the hold-out sample and only 25% for the training sample. To finally forecast ŷT+h we first update the parameter posterior distributions by re-estimating the individual models based on the whole data available (i.e. T observa”
Model Averaging and Its Use in Economics (2017, methodology): “(2010)) and the Fortran code used by Fernández et al. (2001a). Recently, a number of R-packages have been created, in particular the frequently used BMS package (Feldkircher and Zeugner, 2014).”
Merging Seasonal Rainfall Forecasts from Multiple Statistical Models through Bayesian Model Averaging (2012, methodology): “Indeed, there is much literature in support of using predictive performance measures for model choice and combination based on the idea that a model is only as good as its predictions (e.g., Eklund and Karlsson 2007; Geweke and Whiteman 2006; Jackson et al. 2009).”
Advances in Forecasting Under Instability (2012, unknown): “A Monte Carlo analysis of the e¤ects of parameter breaks on out-of-sample forecasting performance in BMAs is considered by Eklund and Karlsson (2005).”
Averaging Forecasts from Vars with Uncertain Instabilities (2006, unknown): “(1)In the interest of tractability, we abstract from the predictive likelihood approach to model and forecast averaging examined in Eklund and Karlsson (2007). Our forecasts based on out-of-sample MSE weighting of course incorporate some of the flavor of weighting on the basis of predictive likeliho”
A weight-relaxed model averaging approach for high-dimensional generalized linear models (2017, methodology): “…Hoeting (1997), Hoeting et al. (1999)], focused information criterion (FIC) model averaging [Claeskens and Hjort (2003), Hjort and Claeskens (2003)], Bayesian model averaging using predictive measures [Eklund and Karlsson (2007)], and predictive likelihood model averaging [Ando and Tsay (2010)].”
A Model-Averaging Approach for High-Dimensional Regression (2014, methodology): “…2001), the Mallows Cp model averaging (Hansen 2007; Wan, Zhang, and Zou 2010; see also Mallows 1973), Bayesian model averaging using predictive measures (Eklund and Karlsson 2007), jackknife model averaging (Hansen and Racine 2012), and predictive likelihood model averaging (Ando and Tsay 2010).”
Inverse groundwater modeling for hydraulic conductivity estimation using Bayesian model averaging and variance window (2008, background): “Some studies have suggested the use of subsets of data as training data to estimate prior model parameter distribution functions [Eklund and Karlsson, 2007]; however, using subsets of data to evaluate prior model parameter probabilities is arguable because prior probabilities express someone’s…”
Benchmark Priors Revisited: On Adaptive Shrinkage and the Supermodel Effect in Bayesian Model Averaging (2009, unknown): “The posterior predictive distribution of ŷ is then given as a a multivariate t-distribution of dimension l (Eklund and Karlsson, 2007, equation (A.15))33 ŷ|X̂,X, y, g ∼ tl(ȳ + sX̂β̂,Σ, N − 1) where Σ = ( Il + sX̂(X ′X)−1X̂ ′ ) y′y N − 1 (1− sR2) Here, s denotes the shrinkage factor s = g 1+g ,…”
Generalized Shrinkage Methods for Forecasting Using Many Predictors (2012, background): “…2010; Carriero, Kapetanios, and Marcellino 2011), Bayesian model averaging (BMA; Jacobson and Karlsson 2004; Koop and Potter 2004; Wright 2009; Eklund and Karlsson 2007), bagging (BG; Inoue and Kilian 2008), Lasso (Bai and Ng 2007; De Mol, Giannone, and Reichlin 2008), boosting (Bai and Ng…”