Hi, I have the simulation results of the following structure: run par measured 1 10 12 2 10 14 1 20 20 2 20 26 Where "run" is the simulation run number, "par" is the parameter of the simulation, and "measured" is the value measured in the simulation. This is only a simple example of my results. There are many values measured and many parameters. But the basic structure stays the same: there are many runs (identified by the run number) for the same values of the parameters with various measured values -- they constitute a sample. I would like to calculate the mean of the "measured" value for a sample, and so I would like to obtain the output as follows: par mean 10 13 20 23 I would appreciate it if someone could write me how to do it. Thank you, Irek
On 17.01.2012 12:31, Irek Szczesniak wrote:> Hi, > > I have the simulation results of the following structure: > > run par measured > 1 10 12 > 2 10 14 > 1 20 20 > 2 20 26 > > Where "run" is the simulation run number, "par" is the parameter of > the simulation, and "measured" is the value measured in the > simulation. This is only a simple example of my results. There are > many values measured and many parameters. But the basic structure > stays the same: there are many runs (identified by the run number) for > the same values of the parameters with various measured values -- they > constitute a sample. > > I would like to calculate the mean of the "measured" value for a > sample, and so I would like to obtain the output as follows: > > par mean > 10 13 > 20 23 > > I would appreciate it if someone could write me how to do it.For you data in a data.frame called dat: aggregate(measured ~ par, dat, mean) Uwe Ligges> > Thank you, > Irek > > ______________________________________________ > 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.
Thank you, Uwe, for your help! I have more measurements (m1, m2) and more parameters (par1, par2). I can calculate the means of m1 and m2 this way: aggregate(cbind(m1, m2) ~ par1 + par2, dat, mean) However, I also need to calculate the standard error of the mean, and the variance for the sample, and I would like to have them output as extra columns next to the column with means. Again, I would appreciate any help! On 17.01.2012 15:09, Uwe Ligges wrote:> > > On 17.01.2012 12:31, Irek Szczesniak wrote: >> Hi, >> >> I have the simulation results of the following structure: >> >> run par measured >> 1 10 12 >> 2 10 14 >> 1 20 20 >> 2 20 26 >> >> Where "run" is the simulation run number, "par" is the parameter of >> the simulation, and "measured" is the value measured in the >> simulation. This is only a simple example of my results. There are >> many values measured and many parameters. But the basic structure >> stays the same: there are many runs (identified by the run number) for >> the same values of the parameters with various measured values -- they >> constitute a sample. >> >> I would like to calculate the mean of the "measured" value for a >> sample, and so I would like to obtain the output as follows: >> >> par mean >> 10 13 >> 20 23 >> >> I would appreciate it if someone could write me how to do it. > > > For you data in a data.frame called dat: > > aggregate(measured ~ par, dat, mean) > > Uwe Ligges > > >> >> Thank you, >> Irek >> >> ______________________________________________ >> 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. >-- Ireneusz (Irek) Szczesniak http://www.irkos.org
Try using the function in the plyr package. E.g.,
> z <- data.frame( # your toy dataset
run = c(1, 2, 1, 2),
par = c(10, 10, 20, 20),
measured = c(12, 14, 20, 26))
> library(plyr)
> ddply(z, .(par), summarize, meanMeasured=mean(measured),
sdMeasured=sd(measured))
par meanMeasured sdMeasured
1 10 13 1.414214
2 20 23 4.242641
Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com
> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org] On Behalf Of Ireneusz
> Szczesniak
> Sent: Tuesday, January 17, 2012 2:43 PM
> To: r-help at r-project.org
> Subject: Re: [R] Mean of simulation runs given in a table
>
> Thank you, Uwe, for your help! I have more measurements (m1, m2) and
> more parameters (par1, par2). I can calculate the means of m1 and m2
> this way:
>
> aggregate(cbind(m1, m2) ~ par1 + par2, dat, mean)
>
> However, I also need to calculate the standard error of the mean, and
> the variance for the sample, and I would like to have them output as
> extra columns next to the column with means.
>
> Again, I would appreciate any help!
>
> On 17.01.2012 15:09, Uwe Ligges wrote:
> >
> >
> > On 17.01.2012 12:31, Irek Szczesniak wrote:
> >> Hi,
> >>
> >> I have the simulation results of the following structure:
> >>
> >> run par measured
> >> 1 10 12
> >> 2 10 14
> >> 1 20 20
> >> 2 20 26
> >>
> >> Where "run" is the simulation run number,
"par" is the parameter of
> >> the simulation, and "measured" is the value measured in
the
> >> simulation. This is only a simple example of my results. There are
> >> many values measured and many parameters. But the basic structure
> >> stays the same: there are many runs (identified by the run number)
for
> >> the same values of the parameters with various measured values --
they
> >> constitute a sample.
> >>
> >> I would like to calculate the mean of the "measured"
value for a
> >> sample, and so I would like to obtain the output as follows:
> >>
> >> par mean
> >> 10 13
> >> 20 23
> >>
> >> I would appreciate it if someone could write me how to do it.
> >
> >
> > For you data in a data.frame called dat:
> >
> > aggregate(measured ~ par, dat, mean)
> >
> > Uwe Ligges
> >
> >
> >>
> >> Thank you,
> >> Irek
> >>
> >> ______________________________________________
> >> 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.
> >
>
>
> --
> Ireneusz (Irek) Szczesniak
> http://www.irkos.org
>
> ______________________________________________
> 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.
Thank you, William, for your help! It works great. My final call
looks like this:
pars <- c(.(nodes), .(load), .(buffer), .(deflections))
ddply(i, pars, summarize,
mm_created = mean(mean_created),
ms_created = mean(sdev_created),
mm_admitted = mean(mean_admitted),
ms_admitted = mean(sdev_admitted),
mm_dropped = mean(mean_dropped),
ms_dropped = mean(sdev_dropped),
mm_delivered = mean(mean_delivered),
ms_delivered = mean(sdev_delivered))
2012/1/18 William Dunlap <wdunlap at tibco.com>:> Try using the function in the plyr package. ?E.g.,
> ?> z <- data.frame( # your toy dataset
> ? ? ? run = c(1, 2, 1, 2),
> ? ? ? par = c(10, 10, 20, 20),
> ? ? ? measured = c(12, 14, 20, 26))
> ?> library(plyr)
> ?> ddply(z, .(par), summarize, meanMeasured=mean(measured),
sdMeasured=sd(measured))
> ? ?par meanMeasured sdMeasured
> ?1 ?10 ? ? ? ? ? 13 1.414214
> ?2 ?20 ? ? ? ? ? 23 4.242641
>
> Bill Dunlap
> Spotfire, TIBCO Software
> wdunlap tibco.com
>
>> -----Original Message-----
>> From: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org] On Behalf Of Ireneusz
>> Szczesniak
>> Sent: Tuesday, January 17, 2012 2:43 PM
>> To: r-help at r-project.org
>> Subject: Re: [R] Mean of simulation runs given in a table
>>
>> Thank you, Uwe, for your help! ?I have more measurements (m1, m2) and
>> more parameters (par1, par2). ?I can calculate the means of m1 and m2
>> this way:
>>
>> aggregate(cbind(m1, m2) ~ par1 + par2, dat, mean)
>>
>> However, I also need to calculate the standard error of the mean, and
>> the variance for the sample, and I would like to have them output as
>> extra columns next to the column with means.
>>
>> Again, I would appreciate any help!
>>
>> On 17.01.2012 15:09, Uwe Ligges wrote:
>> >
>> >
>> > On 17.01.2012 12:31, Irek Szczesniak wrote:
>> >> Hi,
>> >>
>> >> I have the simulation results of the following structure:
>> >>
>> >> run par measured
>> >> 1 10 12
>> >> 2 10 14
>> >> 1 20 20
>> >> 2 20 26
>> >>
>> >> Where "run" is the simulation run number,
"par" is the parameter of
>> >> the simulation, and "measured" is the value measured
in the
>> >> simulation. This is only a simple example of my results. There
are
>> >> many values measured and many parameters. But the basic
structure
>> >> stays the same: there are many runs (identified by the run
number) for
>> >> the same values of the parameters with various measured values
-- they
>> >> constitute a sample.
>> >>
>> >> I would like to calculate the mean of the "measured"
value for a
>> >> sample, and so I would like to obtain the output as follows:
>> >>
>> >> par mean
>> >> 10 13
>> >> 20 23
>> >>
>> >> I would appreciate it if someone could write me how to do it.
>> >
>> >
>> > For you data in a data.frame called dat:
>> >
>> > aggregate(measured ~ par, dat, mean)
>> >
>> > Uwe Ligges
>> >
>> >
>> >>
>> >> Thank you,
>> >> Irek
>> >>
>> >> ______________________________________________
>> >> 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.
>> >
>>
>>
>> --
>> Ireneusz (Irek) Szczesniak
>> http://www.irkos.org
>>
>> ______________________________________________
>> 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.