欢迎来到天天文库
浏览记录
ID:24018839
大小:3.89 MB
页数:60页
时间:2018-11-12
《基于独立成分分析的多维高频数据波动分析》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、阵的研究中,构建了具有长记忆性的ICA—ARFIMA模型对多维资产高频收益率序列进行分析。通过ICA处理将多维资产收益率序列转换为几个相互独立的成分,然后分别计算各独:立成分的“己实现”波动率并进行相关模型估计,从而达到简化模型参数估计的目的。通过实证发现独立“已实现”波动率和对数独立“已实现”波动率都具有显著的长记忆性,ICA.ARFIMA模型能够很好地对多维资产收益率的“已实现”协方差矩阵进行估计。此外,我们将ICA—ARFIMA模型估计得到的“已实现”波动率引入到了风险价值VaR的计算中。实证结果表明ICA—ARFIMA模型得到的波动率能够很好地对资产收益率的风险进行刻画,比直
2、接利用原收益率序列估计得到的VaR值有更高的估计精度。关键词:独立成分分析;长记忆性;DCC—GARCH;ICA—GARCH;ICA—ARFIMA;VaRVOLATILITYMODELINGOFHIGH—FREQUENCYINTRADAYDATABASED0NINDEPENDENTCOM
3、PONENTSANALYSISABSTRACTWiththeintegrationoftheworldeconomy,theconnectionbetweenfinancialmarketsbothathomeandabroadisbecomingcloser.Asanimportantcompon
4、entinChina'sfinancialmarket,moreandmoreattentionsarepaidtothestockmarket.Inaddition,thestockmarketisinthedevelopingandimproving,allparticipantsshouldbefullyawareoftheriSl(sofitandhaveastrongawarenessofriskprevention.Volatilitymodelingisveryimportantinthemanagementofrisk.TheGeneralizedAutoregres
5、siveConditionalHeteroskedasticicy(GARCH)modelandStochasticVolatility(sv)modelarethemostimportantvolatilitymodelsoffinancialtimeseriesmodeling.Inthecaseofmultivariate,GARCHmodelisgeneralizedtothemukivariatecase.TheMultivariateGARCHmodelplaysasignificantroleintheassetallocationandportfolioselecti
6、on.Inthispaper,wesuccinctlysummarythedevelopmentoftheMultivariateGARCHmodelthenapplythemtoanalyzethevolatilityofthereturns.WesuggestusingDCC—GARCHmodelwithdynamiccorrelationICA—GARCHmodelbasedontheIndependentComponentsAnalysis(ICA)toanalyzethevolatilitylinksbetweeneightsliver—relatedstockslikeY
7、uguangGold&lead,TonglingNonferrousMetals,Jiangxicopper,YunnanChihongZinc&Germanium,WesternMining,ShengdaMining,ZijinMiningShenzhenZhongjinLingnanbasedonhigh一仔equencyintradaydataoffive-minutereturns.Theexperimentalresultsshowthatthereexistscorrelatioasbetweenthesliver-relatedstocksthecorrelation
8、sarevaryingovertime.WealsofindthatICA—GARCHmodelismoreeffectivetomodelmultivariatetimeseriesthanDCC—GARCHmodelbycomparingtheautocorrelationofthetwomodels’residualsICA—GARCHmodelcostsleSStimethantheDCC.GARCHmodel.Meanwhile,weintrod
此文档下载收益归作者所有