资源描述:
《自适应的投资组合交易系统外文文献翻译》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、外文文献翻译原文及译文标题:Anadaptiveportfoliotradingsystem:Arisk-returnportfoliooptimizationusingrecurrentTeinfbrcementlearningwithexpectedmaximumdrawdown作者:Saud,Steve期干H:ExpertSystemswithApplications,第87卷,267・279页年份:2017原文Anadaptiveportfoliotradingsystem:Arisk-returnportfoliooptimizationusingrecurre
2、ntreinforcementlearningwithexpectedmaximumdrawdownSaud,SteveAbstractDynamiccontroltheoryhaslongbeenusedinsolvingoptimalassetallocationproblems,andanumberoftradingdecisionsystemsbasedonreinforcementlearningmethodshavebeenappliedinassetallocationandportfoliorebalancing.Inthispaper,weextendt
3、heexistingworkinrecurrentreinforcementlearning(RRL)andbuildanoptimalvariableweightportfolioallocationunderacoherentdownsideriskmeasure,theexpectedmaximumdrawdown,E(MDD).Inparticular,weproposearecurrentreinforcementlearningmethod,withacoherentriskadjustedperformanceobjectivefunction,theCal
4、marratio,toobtainbothbuyandsellsignalsandassetallocationweights.Usingaportfolioconsistingofthemostfrequentlytradedexchange-tradedfunds,weshowthattheexpectedmaximumdrawdownriskbasedobjectivefunctionyieldssuperiorreturnperformancecomparedtopreviouslyproposedRRLobjectivefunctions(i.e.theShar
5、peratioandtheSterlingratio),andthatvariableweightRRLlong/shortportfoliosoutperformequalweightRRLlong/shortportfoliosunderdifferenttransactioncostscenarios.WefurtherproposeanadaptiveE(MDD)riskbasedRRLportfoliorebalancingdecisionsystemwithatransactioncostandmarketconditionstop-lossretrainin
6、gmechanism,andweshowthattheproposedportfoliotradingsystemrespondstotransactioncosteffectsbetterandoutperformshedgefundbenchmarksconsistently.Keywords:Recurrentreinforcementlearning,Expectedmaximumdrawdown,Optimalportfoliorebalancing,Downsiderisk1.1ntroductionInfinancialinvesting,ageneralg
7、oalistodynamicallyallocateasetofassetstomaximizethereturnsovertimeandminimizerisksimultaneously.Forinvestorsitisessentialtobeabletoinvestinaportfoliothatcansatisfytheirpresetgoalsbybuildinganoptimalportfolioinitiallyandsubsequentlyrebalancingitoptimally.Portfoliothe