[R] maxNR - Error in p(a, b) : element 1 is empty; the part of the args list of '*' being evaluated was: (b, t)

[Arne Henningsen]

On 22 March 2010 16:31, 4-real <danielkjaer90 at gmail.com> wrote:
>
> Hello everyone...
> We were trying to implement the Newton-Raphson method in R, and estimate the
> parameters a and b, of a function, F, however we can't seem to implement
> this the right way. Hope you can show me the right way to do this. I think
> what we want R to do is to read the data from the website and then peform
> maxNR on the function, F. Btw the version of R being used is "RGui for
> Windows" if it helps to know this.
>
> R-code below:
>
>  library(maxLik)
>> require(maxLik)
>>
>> x <-
>> read.table('http://www.math.ku.dk/kurser/2008-09/blok4/stat2/doku/data/Eksempel_6_3.txt',
>> header = TRUE);
>> t <- log(x$Koncentration);
>> X <- x$Status;
>>
>> p <- function(a,b) exp(a+b*t)/(1+exp(a+b*t));
>> S <- sum(X);
>> SP <- sum(t*X);
>>
>> F <- function(a,b) {
> +     c(sum(p(a,b)) - S,
> +       sum(t*p(a,b)) - SP)
> + }
>>
>>
>> z <- maxNR(F, start=1, print.level=2)
> Error in p(a, b) : element 1 is empty;
>   the part of the args list of '*' being evaluated was:
>   (b, t)

You forgot to provide argument "b" for function "F":

R> z <- maxNR(F, start=1, b=1, print.level=2)
----- Initial parameters: -----
fcn value: -15.42843
    parameter initial gradient free
[1,]         1        -22.49016    1
Condition number of the (active) hessian: 1
-----Iteration 1 -----
-----Iteration 2 -----
-----Iteration 3 -----
-----Iteration 4 -----
-----Iteration 5 -----
-----Iteration 6 -----
--------------
gradient close to zero. May be a solution
6  iterations
estimate: -0.6288598
Function value: -1.536789


/Arne


-- 
Arne Henningsen
http://www.arne-henningsen.name