help on retrieving output from by( ) for regression

Look more carefully at 
?lm 
at the See Also section ...

X <- rnorm(30)
Y <- rnorm(30)
lm(Y~X)
summary(lm(Y~X))

Best,
Matthias
Hi all 

I used a function 
qtrregr <- by(AB, AB$qtr, function(AB) lm(AB$X~AB$Y))
objective is to run a regression on quartery subsets in the 
data set AB, having variables X and Y, grouped by variable qtr.

Now i retrieved the output using qtrregr, however it only 
showed the coefficients (intercept and B) with out 
significant levels and residuals for each qtr. Can some on 
help me on how can retrieve the detailed regression output.

rgds

snvk

______________________________________________
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

Also, looking at the last example in ?by would be helpful:

attach(warpbreaks)
tmp <- by(warpbreaks, tension, function(x) lm(breaks ~ wool, data = x))

# To get coefficients:
sapply(tmp, coef)

# To get residuals:
sapply(tmp, resid)

# To get the model matrix:
sapply(tmp, model.matrix)

To get the summary() output, I suspect that using:

  lapply(tmp, summary)

would yield more familiar output as compared to using:

  sapply(tmp, summary)

The output from the latter might require a bit more "navigation" through
the resultant matrix, depending upon how the output is to be ultimately
used.

HTH,

Marc Schwartz
Look more carefully at 
?lm 
at the See Also section ...

X <- rnorm(30)
Y <- rnorm(30)
lm(Y~X)
summary(lm(Y~X))

Best,
Matthias

Hi all 

I used a function 
qtrregr <- by(AB, AB$qtr, function(AB) lm(AB$X~AB$Y))
objective is to run a regression on quartery subsets in the 
data set AB, having variables X and Y, grouped by variable qtr.

Now i retrieved the output using qtrregr, however it only 
showed the coefficients (intercept and B) with out 
significant levels and residuals for each qtr. Can some on 
help me on how can retrieve the detailed regression output.

rgds

snvk

What about using lmList from the lme4 package?

Randy

Also, looking at the last example in ?by would be helpful:

attach(warpbreaks)
tmp <- by(warpbreaks, tension, function(x) lm(breaks ~ wool, data = x))

# To get coefficients:
sapply(tmp, coef)

# To get residuals:
sapply(tmp, resid)

# To get the model matrix:
sapply(tmp, model.matrix)

To get the summary() output, I suspect that using:

  lapply(tmp, summary)

would yield more familiar output as compared to using:

  sapply(tmp, summary)

The output from the latter might require a bit more "navigation" through
the resultant matrix, depending upon how the output is to be ultimately
used.

HTH,

Marc Schwartz

On Thu, 2005-08-25 at 14:57 +0200, TEMPL Matthias wrote:
Look more carefully at
?lm 
at the See Also section ...

X <- rnorm(30)
Y <- rnorm(30)
lm(Y~X)
summary(lm(Y~X))

Best,
Matthias

Hi all 

I used a function
qtrregr <- by(AB, AB$qtr, function(AB) lm(AB$X~AB$Y))
objective is to run a regression on quartery subsets in the
data set AB, having variables X and Y, grouped by variable qtr.

Now i retrieved the output using qtrregr, however it only
showed the coefficients (intercept and B) with out
significant levels and residuals for each qtr. Can some on
help me on how can retrieve the detailed regression output.

rgds

snvk

______________________________________________
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

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Randy Johnson
Laboratory of Genomic Diversity
NCI-Frederick
Bldg 560, Rm 11-85
Frederick, MD 21702
(301)846-1304
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
That works also (using the example in ?lmList)

library(lme4)

?lmList

fm1 <- lmList(breaks ~ wool | tension, warpbreaks)

However, one still would need to use either sapply() or lapply() as
below to get the details that Krishna is looking for. 

'fm1' above is a list of models (S4 class 'lmList') not overly different
from 'tmp' below, which is a list of models (S3 class 'by').

If you review str(fm1) and str(tmp), you would note that they are
virtually identical, save the use of slots, etc.

HTH,

Marc Schwartz
What about using lmList from the lme4 package?

Randy

On 8/25/05 9:44 AM, "Marc Schwartz" <MSchwartz at mn.rr.com> wrote:

Also, looking at the last example in ?by would be helpful:

attach(warpbreaks)
tmp <- by(warpbreaks, tension, function(x) lm(breaks ~ wool, data = x))

# To get coefficients:
sapply(tmp, coef)

# To get residuals:
sapply(tmp, resid)

# To get the model matrix:
sapply(tmp, model.matrix)

To get the summary() output, I suspect that using:

  lapply(tmp, summary)

would yield more familiar output as compared to using:

  sapply(tmp, summary)

The output from the latter might require a bit more "navigation" through
the resultant matrix, depending upon how the output is to be ultimately
used.

HTH,

Marc Schwartz

On Thu, 2005-08-25 at 14:57 +0200, TEMPL Matthias wrote:
Look more carefully at
?lm 
at the See Also section ...

X <- rnorm(30)
Y <- rnorm(30)
lm(Y~X)
summary(lm(Y~X))

Best,
Matthias

Hi all 

I used a function
qtrregr <- by(AB, AB$qtr, function(AB) lm(AB$X~AB$Y))
objective is to run a regression on quartery subsets in the
data set AB, having variables X and Y, grouped by variable qtr.

Now i retrieved the output using qtrregr, however it only
showed the coefficients (intercept and B) with out
significant levels and residuals for each qtr. Can some on
help me on how can retrieve the detailed regression output.

rgds

snvk

______________________________________________
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

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Randy Johnson
Laboratory of Genomic Diversity
NCI-Frederick
Bldg 560, Rm 11-85
Frederick, MD 21702
(301)846-1304
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

______________________________________________
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
Sorry to reply to my own post, but in reviewing the NAMESPACE file for
lme4, it looks like Doug is perhaps planning to add additional model
object accessor methods for the lmList class, including resid() and
summary(), which are commented out now. coef() is available presently. 

So when in place, it would appear that the use of lmList would be
advantageous over the use of by().

Marc
That works also (using the example in ?lmList)

library(lme4)

?lmList

fm1 <- lmList(breaks ~ wool | tension, warpbreaks)

However, one still would need to use either sapply() or lapply() as
below to get the details that Krishna is looking for. 

'fm1' above is a list of models (S4 class 'lmList') not overly different
from 'tmp' below, which is a list of models (S3 class 'by').

If you review str(fm1) and str(tmp), you would note that they are
virtually identical, save the use of slots, etc.

HTH,

Marc Schwartz

On Thu, 2005-08-25 at 12:17 -0400, Randy Johnson wrote:
What about using lmList from the lme4 package?

Randy

On 8/25/05 9:44 AM, "Marc Schwartz" <MSchwartz at mn.rr.com> wrote:

Also, looking at the last example in ?by would be helpful:

attach(warpbreaks)
tmp <- by(warpbreaks, tension, function(x) lm(breaks ~ wool, data = x))

# To get coefficients:
sapply(tmp, coef)

# To get residuals:
sapply(tmp, resid)

# To get the model matrix:
sapply(tmp, model.matrix)

To get the summary() output, I suspect that using:

  lapply(tmp, summary)

would yield more familiar output as compared to using:

  sapply(tmp, summary)

The output from the latter might require a bit more "navigation" through
the resultant matrix, depending upon how the output is to be ultimately
used.

HTH,

Marc Schwartz

On Thu, 2005-08-25 at 14:57 +0200, TEMPL Matthias wrote:
Look more carefully at
?lm 
at the See Also section ...

X <- rnorm(30)
Y <- rnorm(30)
lm(Y~X)
summary(lm(Y~X))

Best,
Matthias

Hi all 

I used a function
qtrregr <- by(AB, AB$qtr, function(AB) lm(AB$X~AB$Y))
objective is to run a regression on quartery subsets in the
data set AB, having variables X and Y, grouped by variable qtr.

Now i retrieved the output using qtrregr, however it only
showed the coefficients (intercept and B) with out
significant levels and residuals for each qtr. Can some on
help me on how can retrieve the detailed regression output.

rgds

snvk

______________________________________________
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

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Randy Johnson
Laboratory of Genomic Diversity
NCI-Frederick
Bldg 560, Rm 11-85
Frederick, MD 21702
(301)846-1304
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

______________________________________________
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