Matlab

Zhiming Shen 1. First of all, use file consequence data to present moment data. Then name the data as USMH. A=[1:12] m=[] for i=1:51 m=[m;A] End m=[m;[1,2]] d=ones(614,1) y=USHM(1:end,1) serialdate=datenum(y,m,d) log(USHM(1:end,2)) names=cellstr([HMd;LOG]) TS=fints(serialdate,[USHM(1:end,2), Log],names,Monthly,house market) autocorr(Log) parcorr(Log) Clearly, the data is not stationary. lets name what happens if we do the first difference. g=Log-lagmatrix(Log,1) autocorr(g(2:end,1)) parcorr(g(2:end,1)) It seems that the ACF tails off at lag 4, so lets try a AR(4) model to model the short-term autocorrelation. stipulation = garchset(R,4,M,0, varianceModel,Constant, Display, off) [Coeff,Errors,LLF,Innovations,Sigmas,Summary] = garchfit(spec,g(2:end,1)) garchdisp(Coeff, Errors) Mean: ARMAX(4,0,0); Variance: Constant Conditional Probability distribution: Gaussian Number of Model Parameters Estimated: 6 ideal T Parameter Value Error Statistic ----------- ----------- ------------ ----------- C -0.0016837 0.0068689 -0.2451 AR(1) 0.18209 0.042103 4.3248 AR(2) 0.0042378 0.041209 0.1028 AR(3) -0.15103 0.

04533 -3.3317 AR(4) -0.2152 0.046865 -4.5919 K 0.020741 0.0010889 19.0478 autocorr(Innovations) parcorr(Inn ovations) at that place seems to be some! ACF coefficients significatly different from zero at lags 8 and 14. We can use the Ljung-Box test statistic  to test the join opening that . [h,pValue,stat,cValue] = lbqtest(Innovations, lags, 8:14, alpha, .01, dof, 2:8) LjungBox_test = dataset({transpose(8:13),lag},{stat, QStat}, {pValue, pValue}) LjungBox_test = lag QStat pValue 8 50.92693 8.7369e-012 9 55.41725 5.5936e-012 10...If you necessitate to get a full essay, order it on our website: OrderCustomPaper.com

If you want to get a full essay, visit our page: write my paper