Joint estimation of the predictive ability of experts using a multi-output Gaussian process
Abstract
A multi-output Gaussian process (GP) is introduced as a model for the joint posterior distribution of the local predictive ability of set of models and/or experts, conditional on a vector of covariates, from historical predictions in the form of log predictive scores. Following a power transformation of the log scores, a GP with Gaussian noise can be used, which allows faster computation by first using Hamiltonian Monte Carlo to sample the hyper-parameters of the GP from a model where the latent GP surface has been marginalized out, and then using these draws to generate draws of joint predictive ability conditional on a new vector of covariates. Linear pools based on learned joint local predictive ability are applied to predict daily bike usage in Washington DC.
Cite
@article{arxiv.2402.07439,
title = {Joint estimation of the predictive ability of experts using a multi-output Gaussian process},
author = {Oscar Oelrich and Mattias Villani},
journal= {arXiv preprint arXiv:2402.07439},
year = {2024}
}
Comments
22 pages, 7 figures. This paper was included in the first author's PhD thesis: Oelrich, O. (2022) 'Learning Local Predictive Accuracy for Expert Evaluation and Forecast Combination' which can be found at https://su.diva-portal.org/smash/record.jsf?pid=diva2:1708601 . arXiv admin note: text overlap with arXiv:2402.02068