Investing Through Economic Cycles with Ensemble Machine Learning Algorithms
Benoit, Sylvain; Raffinot, Thomas (2018), Investing Through Economic Cycles with Ensemble Machine Learning Algorithms. https://basepub.dauphine.fr/handle/123456789/19851
TypeDocument de travail / Working paper
External document linkhttps://papers.ssrn.com/sol3/papers.cfm?abstract_id=2785583
Series titleSSRN Working Paper Series
MetadataShow full item record
Laboratoire d'Economie de Dauphine [LEDa]
Abstract (EN)Ensemble machine learning algorithms (random forest and boosting) are applied to quickly and accurately detect economic turning points in the United States and in the Eurozone over the past three decades. The two key features of those algorithms are their abilities (i) to entertain a large number of predictors and (ii) to perform both variable selection and estimation simultaneously. The real-time ability to nowcast economic turning points is gauged by using investment strategies based on economic regimes induced by our models. When comparing predictive accuracy and profit measures, the model confidence set procedure is applied to avoid data snooping. We show that such investment strategies achieve impressive risk-adjusted returns: timing the market is thus possible.
Subjects / KeywordsRandom Forest; Boosting, Economic cycles; Profit maximization measures; Model Confidence Set; Machine Learning; Turning Points Detection; C53; E32; E37; G1
Showing items related by title and author.
Ensemble learning based on functional connectivity and Riemannian geometry for robust workload estimation Corsi, Marie-Constance; Chevallier, Sylvain; Barthélemy, Quentin; Hoxha, Isabelle; Yger, Florian Communication / Conférence
Marketing employees' learning through contact with marketing service providers (MSPs): analysis and deconstruction of the transmitted knowledge flow. Moraux, Hélène; Volle, Pierre (2015-05) Communication / Conférence
Vareschi, Thomas; Picard, Dominique; Hoffmann, Marc; Delattre, Sylvain (2012) Article accepté pour publication ou publié