GM(1-1)模型中的MATLAB程序.docVIP

。 MATLAB程序如下： a=-0.1062; b=371.6018; X0=[142 340 200 500 900 800 490 980 463 1100]; u=b/a u = -3.4991e+003 v=X0(1)-u v = 3.6411e+003 for n=0:9 X2(n+1)=v*exp(-a*n)+u end X2 = 1.0e+003 * 0.1420 0.5500 1.0036 1.5081 2.0692 2.6931 3.3868 4.1584 5.0163 5.9705 、将进行一次累减得到预测值序列，将 记作，利用MATLAB计算，程序如下： X3(1)=X2(1) X3 = 142 for m=1:9 X3(m+1)=X2(m+1)-X2(m) End X3 = 142.0000 407.9616 453.6714 504.5026 561.0293 623.8894 693.7926 771.5281 857.9734 954.1044 二、模型的检验——后验差检验： X0mean=mean(X0) X0mean = 591.5000 X0std=std(X0) X0std = 333.6516 diata0=abs(X0-X3) diata0 =0 67.9616 253.6714 4.5026 338.9707 176.1106 203.7926 208.4719 394.9734 145.8956 diata0mean = 179.4350 diata0std=std(diata0) diata0std = 131.3251 c=diata0std/X0std c = 0.3936 、计算小残差概率：，记作， S0=0.6745*X0std S0 = 225.0480 e=abs(diata0-diata0mean) e = 179.4350 111.4734 74.2363 174.9324 159.5357 3.3244 24.3575 29.0369 215.5383 33.5394