我用EVIEWS做线性回归，发现数据归一化前和归一化后差别很大，其中自变量和因变量相差10000倍，需要进行归一化吗？请大家指点!


归一化前
   
   
Variable Coefficient Std. Error t-Statistic Prob.  
   
   
C 0.203722 0.009401 21.67066 0.0000
X 2.66E-06 1.95E-06 1.361730 0.1764
   
   
R-squared 0.018758     Mean dependent var 0.213427
Adjusted R-squared 0.008642     S.D. dependent var 0.061268
S.E. of regression 0.061002     Akaike info criterion -2.735819
Sum squared resid 0.360963     Schwarz criterion -2.683392
Log likelihood 137.4230     F-statistic 1.854308
Durbin-Watson stat 0.672178     Prob(F-statistic) 0.176438
   
   


归一化后：

Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	0.263004	0.028736	9.152512	0.0000
X	0.106988	0.073025	1.465078	0.1462
R-squared	0.021870	Mean dependent var		0.295080
Adjusted R-squared	0.011681	S.D. dependent var		0.185333
S.E. of regression	0.184247	Akaike info criterion		-0.524878
Sum squared resid	3.258920	Schwarz criterion		-0.472124
Log likelihood	27.71904	F-statistic		2.146452
Durbin-Watson stat	0.686809	Prob(F-statistic)		0.146167