摘要
Thispaperpresentsamethodofconstructingamixedgraphwhichcanbeusedtoanalyzethecausalityformultivariatetimeseries.Weconstructapartialcorrelationgraphatfirstwhichisanundirectedgraph.Foreveryundirectededgeinthepartialcorrelationgraph,themeasuresoflinearfeedbackbetweentwotimeseriescanhelpusdecideitsdirection,thenweobtainthemixedgraph.Usingthismethod,weconstructamixedgraphforfuturessugarpricesinZhengzhou(ZF),spotsugarpricesinZhengzhou(ZS)andfuturessugarpricesinNewYork(NF).Theresultshowsthatthereisabi-directionalcausalitybetweenZFandZS,anunidirectionalcausalityfromNFtoZF,butnocausalitybetweenNFandZS.更多还原
出版日期
2016年01月11日(中国期刊网平台首次上网日期,不代表论文的发表时间)