R help - Feb 2007 - LM Model

Dear R-Users,

 

How can I put a pre-defined regression model into to an object of class lm
in order to use the predict.lm function.

 

A simplified example:

 

I would normally run a regression analysis on a dataset, 

 
> germany<-lm(RENT~AGE1, in.mi01)
> summary(germany)
 

Call:

lm(formula = RENT ~ AGE1, data = in.mi01)

 

Residuals:

    Min      1Q  Median      3Q     Max 

-12.193  -3.646  -1.009   2.101  49.025 

 

Coefficients:

             Estimate Std. Error t value Pr(>|t|)    

(Intercept) 13.283639   0.063020  210.78   <2e-16 ***

AGE1        -0.091030   0.003231  -28.17   <2e-16 ***

---

Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05
'.' 0.1 ' ' 1

 

Residual standard error: 6.044 on 19802 degrees of freedom

Multiple R-Squared: 0.03854,    Adjusted R-squared: 0.03849 

F-statistic: 793.7 on 1 and 19802 DF,  p-value: < 2.2e-16

 

and then use the predict.lm function which uses the above specified
regression equation: RENT=13.283639-0.091030*AGE1 

 

But I want to skip the lm function and specify my own regression equation
RENT= 15 -0.15*AGE1 and then use the predict.lm function. However, in order
to use the predict.lm function I need an object of class lm. Is there any
way to do so? Or maybe somebody has another solution? 

 

Thanks in advance,

 

Simon

 

 


	[[alternative HTML version deleted]]

>>> "Simon P. Kempf" <simon.kempf at web.de> 2/9/2007
1:36 am >>> asked
<<<< 
But I want to skip the lm function and specify my own regression
equation RENT= 15 -0.15*AGE1 and then use the predict.lm function.
However, in order to use the predict.lm function I need an object of
class lm. Is there any way to do so? Or maybe somebody has another
solution? 
>>>
Maybe I am missing something, but why not just write the formula
without the predict.lm?

e.g
<<<
AGE1 <- c(18, 19, 20, 21)
RENT= 15 -0.15*AGE1
RENT
>>>
Peter


 

Thanks in advance,

 

Simon

 

 


	[[alternative HTML version deleted]]

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html 
and provide commented, minimal, self-contained, reproducible code.

Simon P. Kempf <simon.kempf <at> web.de> writes:

[SNIP]
> 
> But I want to skip the lm function and specify my own regression equation
> RENT= 15 -0.15*AGE1 and then use the predict.lm function. However, in order
> to use the predict.lm function I need an object of class lm. Is there any
> way to do so? Or maybe somebody has another solution? 
> 
> Thanks in advance,
> 
> Simon
Here is one way. Take a look at help(offset), help(lm), and
help(lm.predict).
> xx <- runif(30)
> yy <- rnorm(30)
> mydata<-data.frame(xx,yy)
> lm(yy~offset(15*rep(1,30))+offset(-0.15*xx)-1,mydata)

Mark Lyman