nls problem

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

nls problem

Prof J C Nash (U30A)
By interpreting the code line by line and looking at the output of the lines, I got the
following result. It looks like it needs the fifu converted to an expression, then
evaluated. This suggests a workaround, but doesn't answer the underlying question about
whether this is supposed to work this way.

JN


> str(fifu)
 language exp(-k * x)
> fifu2<-as.expression(fifu)
>     fit2b <- nls(y ~ fifu2, data = data, start = c(k = 1))
Error in lhs - rhs : non-numeric argument to binary operator
> fifu2
expression(exp(-k * x))
>     fit2be <- nls(y ~ eval(fifu2), data = data, start = c(k = 1))
> fit2be
Nonlinear regression model
  model:  y ~ eval(fifu2)
   data:  data
k
1
 residual sum-of-squares: 1.604e-06

Number of iterations to convergence: 1
Achieved convergence tolerance: 1.603e-06


--------------------------------------

>
> Message: 13
> Date: Tue, 03 Jul 2012 13:54:11 +0200
> From: "joerg van den hoff" <[hidden email]>
> To: [hidden email]
> Subject: [R] nls problem
> Message-ID: <[hidden email]>
> Content-Type: text/plain; charset=iso-8859-15; format=flowed;
> delsp=yes
>
> hi list,
>
> used versions: 2.12.1 and 2.14.0 under ubuntu and macosx.
>
> I recently stumbled over a problem with `nls', which occurs if the model  
> is not specified explicitly but via an evaluation of a 'call' object.  
> simple example:
>
> 8<--------------------------------------------------------------------------------------
>
> nlsProblem <- function (len = 5) {
> #=======================================================================
>     # purpose: to demonstrate an apparent problem with `nls' which occurs,
>     # if the model is specified by passing th lhs as an evaled 'call'
>     # object.  The problem is related to the way `nls' tries to compute
>     # its internal variable `varIndex' which rests on the assumption that
>     # the dependent ("y") and, possibly, the independent ("x") variable
>     # are identified by having a length equal to the `nls' variable
>     # `respLength'. the problem arises when there are `varNames'
>     # components accidentally having this length, too.
>
>     # in the present example, setting the number of data points to
>     # len=2 triggers the error since the `call' object `fifu' has this
>     # length, too and `nls' constructs an erroneous `mf$formula' internally.
> #=======================================================================
>     #generate some data
>     x <- seq(0, 4, len = len)
>     y <- exp(-x)
>     y <- rnorm(y, y, .001*y)
>     data <- list(x = x, y = y)
>
>     #define suitable model
>     model <- y ~ exp(-k*x)
>     fifu  <- model[[3]]
>

This is where my output above should be placed. JN


>     #this fit is fine:
>     fit1 <- nls(model, data = data, start = c(k = 1))
>     print(summary(fit1))
>
>     #this fit crashes `nls' if len = 2:
>     fit2 <- nls(y ~ eval(fifu), data = data, start = c(k = 1))
>     print(summary(fit2))
> }
>
> 8<--------------------------------------------------------------------------------------
>
> to see the problem call `nlsProblem(2)'.
>
> as explained in the above comments in the example function, I tracked it  
> down to the way
> `nls' identifies x and y in the model expression. the problem surfaces in  
> the line
>
>          varIndex <- n%%respLength == 0
>
> (line 70 in the function listing from within R) which, in the case of  
> `fit2' in the above
> example always returns a single TRUE index as long as `len != 2' (which  
> seems fine for the
> further processing) but returns a TRUE value for the index of `fifu' as  
> well if `len == 2'.
>
> question1: I'm rather sure it worked about 6 months ago with (ca.) 2.11.x  
> under ubuntu. have there been changes in this area?
> question2: is something like the `fit2' line in the example expected to  
> work or not?
> qeustion3: if it is not expected to work, should not the manpage include a  
> corresponding caveat?
> question4: is there a a substitute/workaround for the `fit2' line which  
> still allows to specify the rhs of the model via a variable instead of a  
> constant (explicit) expression or function call?
>
> the above  example is of course construed but in my use case I actually  
> need this sort of thing. is there any chance that
> the way `nls' analyzes its `model' argument can be changed to parse the  
> `eval(fifu)' construct correctly in all cases?
>
> since I'm currently not subscribed to the list I'd appreciate if responses  
> could be Cc'ed to me directly.
>
> thanks in advance,
>
> joerg
>
>

______________________________________________
[hidden email] 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.