[R] Why is the program too slow?

[Jian Zhang]

Hi,everyone.
I use the following program calculates Fisher's alpha from counts of 
individuals and species. The program is wrote by Prof. Kyle Harm.
However, when I run the program, it can work very quickly sometimes, but it can 
not work very well sometimes. It depends on the counts of individuals and 
species.
For example,

> calc.alpha(1000,70)
[1] 17.14375
> calc.alpha(10000,70)
[1] 10.15460
> calc.alpha(100,7)
[1] 1.714375

But,
> calc.alpha(1580,30)
> calc.alpha(1000,7)

It is very slow.

So, what is the problem?
Thanks very much.
                                                      Jian Zhang


# The following function calculates Fisher's alpha from counts of individuals 
and species.
# Note that this program assumes that the true value of alpha lies within the 
range 0.001¨C10000
# (a likely assumption for local assemblages of organisms).
# The function returns "-1" if there is a problem.

calc.alpha=function(n.orig, s.orig)
 {
  a=numeric()

  len.n=length(n.orig)
  len.s=length(s.orig) 
   
  if(len.n != len.s)
    { return(-1)  }
   
  for(i in 1:len.n)
   {
     if(n.orig[i]<=0 | s.orig[i]<=0 | n.orig[i]<=s.orig[i])
 { a[i]=(-1) }
     
     else
      {
       low.a=0.001
       high.a=10000
       low.s = low.a*log(1+(n.orig[i]/low.a))
       high.s = high.a*log(1+(n.orig[i]/high.a)) 
       
       if((s.orig[i]<=low.s) | (s.orig[i]>=high.s))
        { a[i]=(-1) }
       
       else
        {
         use.s=s.orig[i]+1 
         while(s.orig[i] != use.s)
          {
           use.a=(low.a+high.a)/2
           use.s=use.a*log(1+(n.orig[i]/use.a))
           
           if(s.orig[i]<use.s)
  { high.a=use.a }

           if(s.orig[i]>use.s)
  { low.a=use.a }
          }
         a[i]=use.a 
        }
      }
   }

   return(a)
 }

        [[alternative HTML version deleted]]


______________________________________________
R-help@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.