qiantehao 发表于 2012-3-22 21:52
我都按照那个步骤把文件放好了,运行Krolzig的msvecm模型的结果如下:不知道这个结果代表什么,希望老师给 ...
要作跟9.3 A Markov-switching vector equilibrium correction model相同的结果
我已帮你修改程序
请将底下copy,存成 xxx.ox
#include <oxstd.h>
#import <msvar130>
main()
{
decl time=timer();
decl msvar = new MSVAR();
msvar->IsOxPack(FALSE);
// MSVAR settings (TRUE: automatic StdErrors, TRUE: automatic DrawResults, TRUE: save gwg files)
msvar->SetOptions(TRUE,TRUE,TRUE);
// Load data and select variables
msvar->Load("kroto.xls");
msvar->SetPrint(TRUE,TRUE); // all results are printed
//(tolerance, max.#iterations, max.#iterations for MSteps)
decl M=3; // number of regimes
decl p=1; // number of lages
msvar->Select(Y_VAR, { "DN", 0, p, "DY", 0, p}); //endogenous
msvar->Select(X_VAR, { "Cyn", 1, 1}); //exogenous
msvar->SetSample(1962,1,1997,1);
// Specify the MS-VAR
msvar->SetModel(MSIH, M); // model specification
// ESTIMATE
msvar->Estimate(); // estimates
// Graphics
msvar->DrawResults(); // shows graphics
msvar->DrawErrors(TRUE); // shows graphics
msvar->DrawFit(); // shows graphics
msvar->StdErr(); // calculates standard errors
delete msvar;
}
////////Results:
---------- Calculate starting values ---------------
It. 0 LogLik = -129.2811 Pct.Change =100.0000
It. 1 LogLik = -119.2110 Pct.Change = 7.7893
It. 2 LogLik = -115.4770 Pct.Change = 3.1323
It. 3 LogLik = -114.4653 Pct.Change = 0.8761
.....
.....
It. 34 LogLik = -112.8070 Pct.Change = 0.0002
It. 35 LogLik = -112.8069 Pct.Change = 0.0001
It. 36 LogLik = -112.8068 Pct.Change = 0.0001
---------- EM algorithm converged -----------------
EQ( 1) MSIH(3)-VARX(1) model of (DN,DY)
Estimation sample: 1962 (3) - 1997 (1)
no. obs. per eq. : 139 in the system : 278
no. parameters : 27 linear system : 11
no. restrictions : 10
no. nuisance p. : 6
log-likelihood : -112.8068 linear system : -145.4374
AIC criterion : 2.0116 linear system : 2.2509
HQ criterion : 2.2432 linear system : 2.3453
SC criterion : 2.5816 linear system : 2.4831
LR linearity test: 65.2612 Chi(10) =[0.0000] ** Chi(16)=[0.0000] ** DAVIES=[0.0000] **
---------- matrix of transition probabilities ------
Regime 1 Regime 2 Regime 3
Regime 1 0.8201 0.0364 0.1435
Regime 2 0.0532 0.9435 0.0033
Regime 3 0.0350 0.0738 0.8911
---------- regime properties ----------------------
nObs Prob. Duration
Regime 1 29.1 0.2058 5.56
Regime 2 64.7 0.5072 17.68
Regime 3 45.2 0.2869 9.19
---------- coefficients ----------------------------
DN DY
Const(Reg.1) -0.035353 -0.225165
Const(Reg.2) 0.233379 0.604256
Const(Reg.3) 0.431374 1.181520
DN_1 0.538135 0.145198
DY_1 0.023442 0.001669
Cyn_1 0.071826 -0.022038
SE (Reg.1) 0.450898 1.036841
SE (Reg.2) 0.136691 0.427576
SE (Reg.3) 0.289862 0.783621
---------- contemporaneous correlation -------------
Regime 1
DN DY
DN 1.0000 0.8474
DY 0.8474 1.0000
Regime 2
DN DY
DN 1.0000 0.3916
DY 0.3916 1.0000
Regime 3
DN DY
DN 1.0000 0.6685
DY 0.6685 1.0000
---------- standard errors -------------------------
DN DY
Const(Reg.1) 0.1113 0.2786
Const(Reg.2) 0.0433 0.1147
Const(Reg.3) 0.0863 0.2159
DN_1 0.0596 0.1689
DY_1 0.0363 0.0978
Cyn_1 0.0226 0.0667
---------- t - values ------------------------------
DN DY
Const(Reg.1) -0.3176 -0.8083
Const(Reg.2) 5.3955 5.2692
Const(Reg.3) 4.9963 5.4736
DN_1 9.0319 0.8597
DY_1 0.6460 0.0171
Cyn_1 3.1774 -0.3304