Learning Time-Varying Forecast Combinations. Mandel, A. & Sani, A. Working Paper Centre d'Economie de la Sorbonne 2016.36, 2016. ![Learning Time-Varying Forecast Combinations [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex Non-parametric forecast combination methods choose a set of static weights to combine over candidate forecasts as opposed to traditional forecasting approaches, such as ordinary least squares, that combine over information (e.g. exogenous variables). While they are robust to noise, structural breaks, inconsistent predictors and changing dynamics in the target variable, sophisticated combination methods fail to outperform the simple mean. Time-varying weights have been suggested as a way forward. Here we address the challenge to “develop methods better geared to the intermittent and evolving nature of predictive relations” in Stock and Watson (2001) and propose a data driven machine learning approach to learn time-varying forecast combinations for output, inflation or any macroeconomic time series of interest. Further, the proposed procedure “hedges” combination weights against poor performance to the mean, while optimizing weights to minimize the performance gap to the best candidate forecast in hindsight. Theoretical results are reported along with empirical performance on a standard macroeconomic dataset for predicting output and inflation.
@article{Mandel2016b,
abstract = {Non-parametric forecast combination methods choose a set of static weights to combine over candidate forecasts as opposed to traditional forecasting approaches, such as ordinary least squares, that combine over information (e.g. exogenous variables). While they are robust to noise, structural breaks, inconsistent predictors and changing dynamics in the target variable, sophisticated combination methods fail to outperform the simple mean. Time-varying weights have been suggested as a way forward. Here we address the challenge to “develop methods better geared to the intermittent and evolving nature of predictive relations” in Stock and Watson (2001) and propose a data driven machine learning approach to learn time-varying forecast combinations for output, inflation or any macroeconomic time series of interest. Further, the proposed procedure “hedges” combination weights against poor performance to the mean, while optimizing weights to minimize the performance gap to the best candidate forecast in hindsight. Theoretical results are reported along with empirical performance on a standard macroeconomic dataset for predicting output and inflation.},
author = {Mandel, Antoine and Sani, Amir},
journal = {Working Paper Centre d'Economie de la Sorbonne 2016.36},
keywords = {DOLFINS{\_}T1.3,DOLFINS{\_}T2.2,DOLFINS{\_}T2.3,DOLFINS{\_}WP1,DOLFINS{\_}WP2,DOLFINS{\_}working{\_}paper},
mendeley-tags = {DOLFINS{\_}T1.3,DOLFINS{\_}T2.2,DOLFINS{\_}T2.3,DOLFINS{\_}WP1,DOLFINS{\_}WP2,DOLFINS{\_}working{\_}paper},
title = {{Learning Time-Varying Forecast Combinations}},
url = {https://halshs.archives-ouvertes.fr/halshs-01317974/},
year = {2016}
}
Downloads: 0
{"_id":"D4wGxR6QKMDpNMk7Z","bibbaseid":"mandel-sani-learningtimevaryingforecastcombinations-2016","downloads":0,"creationDate":"2016-11-22T10:32:19.137Z","title":"Learning Time-Varying Forecast Combinations","author_short":["Mandel, A.","Sani, A."],"year":2016,"bibtype":"article","biburl":"https://simpolproject.eu/download/Bibliography/DOLFINS2018.06.11v1.bib","bibdata":{"bibtype":"article","type":"article","abstract":"Non-parametric forecast combination methods choose a set of static weights to combine over candidate forecasts as opposed to traditional forecasting approaches, such as ordinary least squares, that combine over information (e.g. exogenous variables). While they are robust to noise, structural breaks, inconsistent predictors and changing dynamics in the target variable, sophisticated combination methods fail to outperform the simple mean. Time-varying weights have been suggested as a way forward. Here we address the challenge to “develop methods better geared to the intermittent and evolving nature of predictive relations” in Stock and Watson (2001) and propose a data driven machine learning approach to learn time-varying forecast combinations for output, inflation or any macroeconomic time series of interest. Further, the proposed procedure “hedges” combination weights against poor performance to the mean, while optimizing weights to minimize the performance gap to the best candidate forecast in hindsight. Theoretical results are reported along with empirical performance on a standard macroeconomic dataset for predicting output and inflation.","author":[{"propositions":[],"lastnames":["Mandel"],"firstnames":["Antoine"],"suffixes":[]},{"propositions":[],"lastnames":["Sani"],"firstnames":["Amir"],"suffixes":[]}],"journal":"Working Paper Centre d'Economie de la Sorbonne 2016.36","keywords":"DOLFINS_T1.3,DOLFINS_T2.2,DOLFINS_T2.3,DOLFINS_WP1,DOLFINS_WP2,DOLFINS_working_paper","mendeley-tags":"DOLFINS_T1.3,DOLFINS_T2.2,DOLFINS_T2.3,DOLFINS_WP1,DOLFINS_WP2,DOLFINS_working_paper","title":"Learning Time-Varying Forecast Combinations","url":"https://halshs.archives-ouvertes.fr/halshs-01317974/","year":"2016","bibtex":"@article{Mandel2016b,\nabstract = {Non-parametric forecast combination methods choose a set of static weights to combine over candidate forecasts as opposed to traditional forecasting approaches, such as ordinary least squares, that combine over information (e.g. exogenous variables). While they are robust to noise, structural breaks, inconsistent predictors and changing dynamics in the target variable, sophisticated combination methods fail to outperform the simple mean. Time-varying weights have been suggested as a way forward. Here we address the challenge to “develop methods better geared to the intermittent and evolving nature of predictive relations” in Stock and Watson (2001) and propose a data driven machine learning approach to learn time-varying forecast combinations for output, inflation or any macroeconomic time series of interest. Further, the proposed procedure “hedges” combination weights against poor performance to the mean, while optimizing weights to minimize the performance gap to the best candidate forecast in hindsight. Theoretical results are reported along with empirical performance on a standard macroeconomic dataset for predicting output and inflation.},\nauthor = {Mandel, Antoine and Sani, Amir},\njournal = {Working Paper Centre d'Economie de la Sorbonne 2016.36},\nkeywords = {DOLFINS{\\_}T1.3,DOLFINS{\\_}T2.2,DOLFINS{\\_}T2.3,DOLFINS{\\_}WP1,DOLFINS{\\_}WP2,DOLFINS{\\_}working{\\_}paper},\nmendeley-tags = {DOLFINS{\\_}T1.3,DOLFINS{\\_}T2.2,DOLFINS{\\_}T2.3,DOLFINS{\\_}WP1,DOLFINS{\\_}WP2,DOLFINS{\\_}working{\\_}paper},\ntitle = {{Learning Time-Varying Forecast Combinations}},\nurl = {https://halshs.archives-ouvertes.fr/halshs-01317974/},\nyear = {2016}\n}\n","author_short":["Mandel, A.","Sani, A."],"key":"Mandel2016b","id":"Mandel2016b","bibbaseid":"mandel-sani-learningtimevaryingforecastcombinations-2016","role":"author","urls":{"Paper":"https://halshs.archives-ouvertes.fr/halshs-01317974/"},"keyword":["DOLFINS_T1.3","DOLFINS_T2.2","DOLFINS_T2.3","DOLFINS_WP1","DOLFINS_WP2","DOLFINS_working_paper"],"downloads":0},"search_terms":["learning","time","varying","forecast","combinations","mandel","sani"],"keywords":["apprentissage statistique","combinaison de prédicteurs","econometrics","econométrie","forecast combination puzzle","forecast combinations","forecasting","machine learning","dolfins_t1.3","dolfins_t2.2","dolfins_t2.3","dolfins_wp1","dolfins_wp2","dolfins_working_paper"],"authorIDs":[],"dataSources":["JBCT4AXyeopWLaLMW"]}