>
y=rnorm(100)
>
x1=rnorm(100)
>
x2=rnorm(100)
>
lm.reg=lm(y~x1+x2,data=data.frame(y,x1,x2)) #data.frame
保证了y=x1+x2的正确性
> lm.reg
Call:
lm(formula = y ~ x1 + x2, data = data.frame(y, x1, x2))
Coefficients:
(Intercept)
#y=0.03493-0.07988x1-0.06208x2
>
summary(lm.reg)
Call:
lm(formula = y ~ x1 + x2)
Residuals:
-2.97055 -0.66086
Coefficients:
(Intercept)
x1
x2
---
Signif. codes:
Residual standard error: 0.9541 on 97 degrees of freedom
Multiple R-squared: 0.007451,
F-statistic: 0.3641 on 2 and 97 DF,
> lm.stp=step(lm.reg)
Start:
y ~ x1 + x2
- x2
<none>
- x1
Step:
y ~ x1
<none>
- x1
Call:
lm(formula = y ~ x1)
Coefficients:
(Intercept)
> plot(lm.reg)
#见图1-4
>
lm.pred=predict(lm.reg,data.frame(x1=1,x2=2),interval="prediction",level=0.95)
> lm.pred
1 -0.1690997 -2.141884 1.803685
#-0.1690997是最后的值,-2.141884~1.803685是置信区间
图1:x轴预测值;y轴残差;异常值是87、56和1样本:反映预测值和真实值的距离;
|
|