MATLAB在金融时间序列中应用的一篇论文

CN4321258/TPISSN10072130XCOMPUTERENGINEERING&SCIENCE2004267Vol126,No17,2004:10072130X(2004)0720100205MATLABXApplicationofMatlabinAnalysingandModelingFinancialTimeSequences1,2LIXing2xu1,CUIJian2fu2(1.,650221;2.,650091)(1.YunnanUniversityofFinanceandEconomics,Kunming650221;2.DepartmentofStatistics,YunnanUniversity,Kunming650091,China):MATLAB,MATLABAbstract:MATLABisanoutstandingmathematicalcomputingtool.Inthispaper,weexpatiatehowtoanalyzeandmodelfinancialtimesequenceswithMATLAB.:;MATLAB;;ARMA;GARCHKeywords:financialtimesequence;MATLAB;technicalanalysis;ARMA;GARCH:O245:A1MATLABMathWorks,,,,MATLAB,,,EviewsTSP,,,MAT2LAB,GARCH,,MATLAB,MATLABGARCH,ARMAXGARCH;2001X:2003204223;:2003206220:(2000A0003M):(1966-),,,,;,,:650221;Tel:[phone];E2mail:[e-mail]:SchoolofStatisticsandInformation,YunnanUniversityofFinanceandEconomics,Kunming,Yunnan650221,P.R.China,,,MATLAB2.1(MACD)MACD(DIF)(DEA),DIF,DEADIFMACD,DIFDEA,;DIFDEA,DIF:macdts=macd(tsobj,series-name);:tsobj,series-name;:IBM10/01/95212/31/95;macd-ibm=macd(part-ibm);plot(macd-ibm);11IBM2.2(Williams%R)Williams%R,-100%0%,Williams%R-20%,,;Williams%R-80%,,:wpctrts=willpctr(tsobj,nperiods);:nperiods(,14);:IBMWilliams%R:wpctr-ibm=willpctr(part-ibm);plot(wpctr-ibm);datetick(x,mm/dd/yy);22.3(RSI)RSI,:rsits=rsindex(tsobj,nperiods,ParameterName,Parameter2Value);:nperiods;:IBMRSI;rsi-ibm=rsindex(part-ibm);plot(rsi-ibm);32IBM3IBM3y(1),y(2),,y(T),{y(t)}t1,,3.13.1.1(AutocorrelationFunction,ACF):k=Cov[y(t),y(t-k)][V(y(t))V(y(t-k))]1/2,,,:[ACF,Lags,Bounds]=autocorr(Series,nLags,M,nSTDs):Series,nLags()ACF,M()LagsM,ACF0,nSTDs()ACFACF,Lags,BoundsACF:szzs.txt,2000szzs=ascii2fts(szzs.txt,1,3);part-szzs=szzs(01/04/2000::12/24/2000);closepri=fts2mtx(part-szzs.close);[ACF,Lags,Bounds]=autocorr(closepri,(length(closepri)-1));autocorr(closepri,(length(closepri)-1));:Bounds=(-0.1307,0.1307),ACFLag410143.1.2(PartialACF,PACF)kAR(k):y(t)=k1y(t-1)++kky(t-k)+tkk,kkYule2Walker,:[PartialACF,Lags,Bounds]=parcorr(Series,nLags,R,nSTDs);:ACF,PartialACF;:2000:parcorr(closepri,(length(closepri)-1));553.2na[AR(na)]:A(q)y(t)=e(t),A(q)=1+a1q-1+a2q-2++anaq-na,|ai|1,q,q-m=(),q-my(t)=y(t-m),e(t)WN(0,2),{e(t)},nc[MA(nc)]:y(t)=C(q)e(t),C(q)=1+c1q-1++cncq-ncAR(na)MA(nc),AR2MA(na,nc):A(q)y(t)=C(q)e(t)11AR(na)MA(nc)ARMA(na,nc)1,MATLABARMA(na,nc)::m=armax(data,orders):data,orders=[nanc],nanc;na=0,MA(nc),nc=0,AR(na)m,FPEAkaike:2000ACF(4)PACF(5),closepri,ACFPACF:logclose=log(closepri);subplot(2,1,1);autocorr(logclose);subplot(2,1,2);parcorr(logclose);6log(closepri)ACFPACF,AR(na),2,,ARIMA(2,0)M=armax(logclose,[20]):A(q)=1-1.147q-1+0.1468q-26closepriACFPACF3.3ARCH(q):y(t)=+tUt,Ut,t,E(Ut)=0,V(ut)=1,t=(0+qi=1ai(y(t-i)-u)2)1/2,ARCH,,ARCH(GeneralizedARCH,GARCH)(Bollerselv,1986,1988)GARCH(p,q):2t=0+qi=1ai2t-i+pj=1j2t-j201,p0j0,1jp,,,GARCH,GARCH(FatTails)(VolatilityClustering)MATLABGARCHGARCH2000,:(1)returnmtx=price2ret(closepri);%Createthedailyreturnvector:returnmtxsubplot(2,2,1);plot(returnmtx);title(TheDailyReturn);subplot(2,2,2);autocorr(returnmtx);title(ACFofDailyRe2turn);subplot(2,2,3);parcorr(returnmtx);title(PACFoftheDailyRe2turn);subplot(2,2,4);autocorr(returnmtx.^2);title(ACFoftheSquareofDailyReturn);720007,,ACFPACF,ACF,Q2ARCH:[H,pValue,Stat,CriticalValue]=lbqtest(returnmtx2mean(return2mtx),[101520],0.05)[H,pValue,Stat,CriticalValue]=archtest(returnmtx2mean(return2mtx),[101520],0.05)Q2:00.071417.134118.307000.193219.469424.995800.058030.788631.4104ARCH:1.00.001528.586818.30701.00.010230.505024.99581.00.018535.319031.4104Q2H=0,;ARCHH=1,ARCH(2)[coeff,errors,LLF,innovations,sigma,summary]=garchfit(re2turnmtx);garchdisp(coeff,errors)GARCH(1,1),:y(t)=0.0018753+t,2t-2.0624e-005+0.505952t-1+0.494052t-1(3),garchfit(LLF)(innova2tions)(sigma),garchplot:garchplot(innovations,sigma)8,,:plot(innovations./sigma),,,989Q2ARCH:[H,pValue,Stat,CriticalValue]=lbqtest((innovations./sigma).^2,[101520],0.05)[H,pValue,Stat,CriticalValue]=archtest(innovations./sigma,[101520],0.05);Q2:00.98083.027618.307000.99734.113224.995800.98978.300031.4104ARCH:00.97083.386018.307000.99574.487024.995800.99557.310931.4104H=0,GARCH(1,1),ARCH4,MATLAB,301,,MATLAB,(),,MATLAB,:[1][]TCMills...2[M].:,2002.[2].MATLAB6.0[M].:,2001.[3][]GEPBox,..:[M].:,1997.(99)(v€c)AP1(t1,s1,m)|t(t)MS(t,s)|(Šmrmovem(smove,t)AP2(t,s2,m)t(v€c)AP1(t1,s1,m)|MS(t,s)|(Šmrmovem(smove,t)AP2(t,s2,m)(v€c)Šm(smove,t)IdentAP1(s1,m)|MS(t,s)|m(smove,t)AP2(t,s2,m)(v€c)IdentAP1(s1,m)|MS(t2,s)|AP2(t2,s2,m)System2,,232Handoff3Handoff2MSAP,;,MSs,(select=true),AP,,3t2;,APAP,t1m35,,802.11MAC,,,,,:[1]RMilner,JParrow,DWalker.ACalculusofMobileProcesses[A].InformationandComputation[M].1992.[2]DSangiorgi,DWalker.The2Calculus:ATheoryofMobilePro2cesses[M].CambridgeUniversityPress,2001.[3]RMilner.CommunicatingandMobileSystems:The2Calculus[M].CambridgeUniversityPress,1999.[4]MAbadi,ADGordon.ACalculusforCryptographicProtocols:TheSpi2Calculus[A].InformationandComputation[M].1999.[5]LCardelli,SDGordon.MobileAmbient[A].ProcofFoundationsofSoftwareScienceandComputationStructures(ToSSaCS),Euro2p