sc 施瓦茨 matlab,[求助]如何在matlab中求线性回归方程的AIC值

function arma=armaa(x)

clc

GG=load('d:/gr.txt');

KK=GG;

GG=dtrend(GG);

aic=autocorr(GG);%自相关

paic=parcorr(GG);%偏相关

mm=-aic;

nn=-paic;

paic(1)=[];

aic(1)=[];

nn(1)=1;

for i=2:13,

for j=1:i,

dd(j)=nn(j);

end

c=filter(dd,1,GG);

for op=1:i-1,

c(op)=[];

end

canca(i-1)=var(c);

end

for i=1:12,

BIC(i)=log(canca(i))+1/length(KK)*log(length(KK));

AIC(i)=log(canca(i))+2*i/length(KK);

end

aic

paic

canca                %残差平方和

AIC

BIC

输出结果：

aic =

-0.5037

0.1220

-0.2108

0.0796

0.0187

0.1169

-0.2162

0.2517

-0.1932

0.0559

-0.1036

0.0123

0.2180

-0.0591

-0.1431

0.0697

-0.0505

0.0115

0.0312

-0.0766

paic =

-0.5041

-0.1781

-0.3551

-0.2831

-0.1772

0.0360

-0.2209

0.1458

0.0688

0.0009

-0.1723

-0.2912

0.1872

0.1437

-0.2386

0.1059

0.2155

-0.3839

-0.2884

-0.2151

canca =

1.0e+003 *

Columns 1 through 11

2.5172    2.4685    2.3110    2.2903    2.3482    2.3340    2.2696    2.2007    1.9902    2.0228    2.1040

Column 12

2.1688

AIC =

Columns 1 through 11

7.8660    7.8815    7.8507    7.8768    7.9368    7.9659    7.9730    7.9772    7.9118    7.9631    8.0376

Column 12

8.1030

BIC =

Columns 1 through 11

7.9019    7.8823    7.8164    7.8074    7.8323    7.8263    7.7983    7.7675    7.6669    7.6832    7.7225

Column 12

7.7529

对于acf和pacf，跟EVIEWS输出结果一样。

matlab计算ARMA的AIC、BIC