[R] lm: how are polynomial functions interpreted?

[Gabor Grothendieck]

Try this to see the X matrix that its fitting y to:

x <- 1:10
model.matrix(~ poly(x, 2))
model.matrix(~ I(x^2))



On Mon, Jan 12, 2009 at 9:01 PM, Carl Witthoft <carl at witthoft.com> wrote:
> damn.
> My apologies to everyone -- I sent one message and it got destroyed somehow.
>
> Here's the part that was missing, which led to rather a lot of confusion on
> all parts.
>
> The two recent responses to a question about lm suggested
>
> 1)  lm(y~poly(x,2))
>
> 2) lm(y~I(x^2))
>
> So my question was *supposed* to be related to how lm operated differently
> (if at all) on these two different 'versions' of a quadratic fit.  My gut
> reaction was that y~I(x^2) would not be the same as y~f(x) where f(x) is
> a+bx+cx^2 .
>
> So what I was trying to find out was just how lm() deals with various
> definitions of the orthogonal polynomials its presented with.  Another way,
> maybe, to ask, is: how does one specify to fit exactly to
>
>  a + bx +cx^2  vs
>
>  bx + cx^2   vs
>  cx^2
> ?
>
> Thanks and apologies again to all the people who quite properly
> misunderstood what I was harping on due to the munging of my first post.
>
> Carl
>
>
> David Winsemius wrote:
>>
>> On Jan 12, 2009, at 5:57 PM, Carl Witthoft wrote:
>>
>>> Well..... *_* ,
>>>
>>> I think it should have been clear that this was not a question for which
>>> any code exists.  In fact, I gave two very specific examples of function
>>> calls.
>>
>> Huh? I think most of the readers of the list saw a basically empty message
>> body. That was the point of Berry's "[nothing deleted]". If you are under
>> the belief that this is a continuation of an earlier question, then it may
>> not be threading up in the manner you hoped for.
>>
>
> ______________________________________________
> R-help at r-project.org 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.
>