资源描述：
《Multi-Scale Volatility Feature Analysisand Prediction of Gold Price多尺度波动特征分析 黄金价格预测》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、InternationalJournalofInformationTechnology&DecisionMakingVol.16,No.1(2017)205–223°cWorldScienti¯cPublishingCompanyDOI:10.1142/S0219622016500504Multi-ScaleVolatilityFeatureAnalysisandPredictionofGoldPrice,†,‡FenghuaWen*,XinYang*,XuGong*andKinKeungLai§,¶,
2、
3、*Busines
4、sSchoolofCentralSouthUniversityChangsha410081,P.R.China†CenterforComputationalFinanceandEconomicAgents,UniversityofEssexColchesterCO43SQ,UK‡InstituteofFinancial,WenZhouUniversityWenzhou325035,P.R.China§InternationalBusinessSchool,ShaanxiNormalUniversity,Xian,P.R.C
5、hina¶DepartmentofManagementSciencesCityUniversityofHongKongKowloon,HongKong
6、
7、mskklai@cityu.edu.hkPublished5December2016Volatilityofgoldpriceisofgreatsigni¯canceforavoidingtheriskofgoldinvestment.Itisnecessarytounderstandthee®ectofexternaleventsandintrinsicregulari
8、tiestomakeaccuratepricepredictions.Thispaper¯rstcomparedEMDwithCEEMDalgorithm,andtheresults¯ndthatCEEMDalgorithmperformanceisbetterthanthatofEMDinanalysisgoldpricevola-tility.Thenthispaperusesthecomplementaryensembleempiricalmodedecomposition(CEEMD)todecomposetheh
9、istoricalpriceofinternationalgoldintopricecomponentsatdi®erentfrequencies,andextractsashort-term°uctuation,ashockfromsigni¯canteventsandalong-termprice.Inaddition,thispapercombinestheIterativecumulativesumofsquares(ICSS)withChowtesttotestthethreeeventpricesforstru
10、cturalbreaks,andanalyzestheInt.J.Info.Tech.Dec.Mak.2017.16:205-223.Downloadedfromwww.worldscientific.come®ectofexternaleventsonvolatilityofgoldpricebycomparingtheexternaleventswiththetestresultsforstructuralbreaks.Finally,thispaperconstructssupportvectormachine(SV
11、M)modelsandarti¯cialneuralnetwork(ANN)onthreeseriesforprediction,and¯ndsthattheby61.174.131.57on08/12/18.Re-useanddistributionisstrictlynotpermitted,exceptforOpenAccessarticles.SVMperformedbetteringoldpricepredictioninone-step-aheadand¯ve-step-ahead,andwhenwecombi
12、netheSVMsandANNswithpricecomponentstomakepredictions,theerrorofthecombinedpredictionissmallerthanSVMsandANNswithseparatetermsofseriesextracted.Keywords:
