R help - Aug 2009 - manipulating text to generate different formulas to use in nls()

Hello,

In doing a series of non-linear estimations of a function which is a sum of a
varying number
of sinusoids, I would like to "autogenerate" the arguments needed by
nls() depending on that
number.

For example, when there are two sinusoids: 

  > nls( y ~ mu + A1 * cos(2*pi*f1*x - P1) + A2 * cos(2*pi*f2*x - P2), data =
some.xy.data,
            start = list( mu=some.value0, A1=some.value1, P1=some.value2,
A2=some.value3,
                         P2=some.value4, f1=some.value5, f2=some.value6  )  )

 Adding a third sinusoid, the model and the starting parameters needed become:

  > nls( y ~ mu + A1 * cos(2*pi*f1*x - P1) + A2 * cos(2*pi*f2*x - P2) + A3 *
cos(2*pi*f3*x - P3),
            data = some.xy.data, start = list( mu=some.value0, A1=some.value1,
            P1=some.value2, A2=some.value3, P2=some.value4, A3=some.value5,
            P3=some.value6, f1=some.value7, f2=some.value8, f3=some.value9 )


So far, I managed to create the "building blocks" I need but when it
comes to defining their
number I still haven't found a solution (see the following code).

##  some fake starting values
par.ini <- data.frame(period=1:5, invR=1:5, a0=1:5, Amplitude=1:5, Phase=1:5)

  n <- nrow(par.ini) - 1                        # I'm not interested in
the last harmonic
  inip <- par.ini[1:n, c(1, 3:5)]             # neither in the second column
    amps <- paste("A", 1:n, sep="")
    phas <- paste("P", 1:n, sep="")
    freqs <- paste("f", 1:n, sep="")

 ( premodl <- paste(amps, " * cos(2*pi*", freqs, "*x - ",
phas, ")", sep="") )


## In the two following instructions, "modl" and "initP" are
defined by hand. This is what I
want to generate "automagically" depending on the number of harmonics
(n).

 ( modl <- paste("y ~ mu", premodl[1],premodl[2], premodl[3],
premodl[4], sep=" + ") )

 ( initP <- list(mu=sum(inip$a0), 
          A1=inip$Amplitude[1], P1=inip$Phase[1], 
          A2=inip$Amplitude[2], P2=inip$Phase[2], 
          A3=inip$Amplitude[3], P3=inip$Phase[3],           
          A4=inip$Amplitude[4], P4=inip$Phase[4], 
          f1=1/inip$period[1], f2=1/inip$period[2],
          f3=1/inip$period[3], f4=1/inip$period[4])  ) 

## not run. These two will be used in
  res <- nls(modl, data=data.frame(y, x), start=initP)
 


Thank you in advance for any insigth.

H?ctor


 
-- 
H?ctor Villalobos <hvillalo at ipn.mx> 
 CICIMAR - IPN
 A.P. 592. Col. Centro 
 La Paz, Baja California Sur, M?XICO. 23000
 Tels. (+52 612) 122 53 44; 123 46 58; 123 47 34  ext. 82425
 Fax.  (+52 612) 122 53 22

Take a look at ?eval

Bart


H?ctor Villalobos-2 wrote:
> 
> Hello,
> 
> In doing a series of non-linear estimations of a function which is a sum
> of a varying number 
> of sinusoids, I would like to "autogenerate" the arguments needed
by nls()
> depending on that 
> number.
> 
> For example, when there are two sinusoids: 
> 
>   > nls( y ~ mu + A1 * cos(2*pi*f1*x - P1) + A2 * cos(2*pi*f2*x - P2),
> data = some.xy.data, 
>             start = list( mu=some.value0, A1=some.value1, P1=some.value2,
> A2=some.value3,
>                          P2=some.value4, f1=some.value5, f2=some.value6  ) 
> )
> 
>  Adding a third sinusoid, the model and the starting parameters needed
> become:
> 
>   > nls( y ~ mu + A1 * cos(2*pi*f1*x - P1) + A2 * cos(2*pi*f2*x - P2) +
A3
> * cos(2*pi*f3*x - P3), 
>             data = some.xy.data, start = list( mu=some.value0,
> A1=some.value1,
>             P1=some.value2, A2=some.value3, P2=some.value4,
> A3=some.value5,
>             P3=some.value6, f1=some.value7, f2=some.value8, f3=some.value9
> )
> 
> 
> So far, I managed to create the "building blocks" I need but when
it comes
> to defining their 
> number I still haven't found a solution (see the following code).
> 
> ##  some fake starting values
> par.ini <- data.frame(period=1:5, invR=1:5, a0=1:5, Amplitude=1:5,
> Phase=1:5)
> 
>   n <- nrow(par.ini) - 1                        # I'm not interested
in
> the last harmonic
>   inip <- par.ini[1:n, c(1, 3:5)]             # neither in the second
> column
>     amps <- paste("A", 1:n, sep="")
>     phas <- paste("P", 1:n, sep="")
>     freqs <- paste("f", 1:n, sep="")
> 
>  ( premodl <- paste(amps, " * cos(2*pi*", freqs, "*x -
", phas, ")",
> sep="") )
> 
> 
> ## In the two following instructions, "modl" and
"initP" are defined by
> hand. This is what I 
> want to generate "automagically" depending on the number of
harmonics (n).
> 
>  ( modl <- paste("y ~ mu", premodl[1],premodl[2], premodl[3],
premodl[4],
> sep=" + ") ) 
> 
>  ( initP <- list(mu=sum(inip$a0), 
>           A1=inip$Amplitude[1], P1=inip$Phase[1], 
>           A2=inip$Amplitude[2], P2=inip$Phase[2], 
>           A3=inip$Amplitude[3], P3=inip$Phase[3],           
>           A4=inip$Amplitude[4], P4=inip$Phase[4], 
>           f1=1/inip$period[1], f2=1/inip$period[2],
>           f3=1/inip$period[3], f4=1/inip$period[4])  ) 
> 
> ## not run. These two will be used in
>   res <- nls(modl, data=data.frame(y, x), start=initP)
>  
> 
> 
> Thank you in advance for any insigth.
> 
> H?ctor
> 
> 
>  
> -- 
> H?ctor Villalobos <hvillalo at ipn.mx> 
>  CICIMAR - IPN
>  A.P. 592. Col. Centro 
>  La Paz, Baja California Sur, M?XICO. 23000
>  Tels. (+52 612) 122 53 44; 123 46 58; 123 47 34  ext. 82425
>  Fax.  (+52 612) 122 53 22
> 
> ______________________________________________
> 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.
> 
> 
-- 
View this message in context:
http://www.nabble.com/manipulating-text-to-generate-different-formulas-to-use-in-nls%28%29-tp24907333p24912932.html
Sent from the R help mailing list archive at Nabble.com.