eviews异方差检验!
eviews异方差检验!
数据如下:
年份 商品房平均销售价格 国民总收入 面积 房地产住宅投资额 城镇人口数
Y X1 X2 X3 X4
1997 1997 78060.8 12464.7 1539.4 39449
1998 2063 83024.3 14125.7 2081.6 41608
1999 2053 88479.2 17640.7 2638.5 43748
2000 2112 98000.5 20603.3 3312.0 45906
2001 2170 108068.2 24625.4 4216.7 48064
2002 2250 119095.7 28524.7 5227.8 50212
2003 2359 135174.0 33774.6 6776.7 52376
2004 2778 159586.7 34677.2 8837.0 54283
2005 3168 184088.6 43682.9 10860.9 56212
2006 3367 213131.7 45471.7 13638.4 57706
2007 3864 251483.2 49831.3 18005.4 59379
最小二乘以后:
Dependent Variable:Y
Method:Least Squares
Date:12/26/09 Time:16:48
Sample:1997 2007
Included observations:11
Variable Coefficient Std.Error t-Statistic Prob.
C 1031.744 61.49013 16.77902 0.0000
X1 0.015024 0.001680 8.940967 0.0000
X2 -0.018366 0.007439 -2.468809 0.0388
R-squared 0.988955 Mean dependent var 2561.909
Adjusted R-squared 0.986194 S.D.dependent var 638.6818
S.E.of regression 75.04552 Akaike info criterion 11.70107
Sum squared resid 45054.64 Schwarz criterion 11.80958
Log likelihood -61.35587 F-statistic 358.1509
Durbin-Watson stat 1.900238 Prob(F-statistic) 0.000000
怎样进行异方差检验及补救 还有自相关检验及补救
进行异方差检验及自相关检验之前已剔除x3 x4
X2的二次项存在异方差,可以用1/X2做加权最小二乘,我试了试可以的,就是输入“ls y/x2 c x1/x2 1/x2 ”自相关是看最后一行Durbin-Watson stat 1.900238,这个统计量接近2说明没有自相关,你做这个没问题.异方差是在菜单中的view-residual test-whitehete……(no cross)