RES: [R] AGREP

[Duncan Murdoch]

On Fri, 13 Feb 2004 14:45:33 +1300 (NZDT), you wrote:

>"Marcos Sanches" <marcos.sanches at ipsos-opinion.com.br> wrote:
>	Ls1<-length(s1)
>	Ls2<-length(s2) 
>	for ( p in 1:ls1){
>	   for (q in 1:ls2){
>	     t1<-levenshteinFast(s1[p],s2[q])
>...
>
>	Ls1=42000
>	Ls2=70000
>	
>	I think I will wait for months untill this program ends. Do you have any
>	sugestion to increase the speed?
>
>The first suggestion has to be "search HARD in the on-line literature;
>as others are bound to have had similar problems."

As I mentioned earlier, R already contains the source to do this using
the same library as agrep() uses.  You just need to write the R
interface to it.

Duncan Murdoch