[R] sampling quadrats of increasing size in a list of matrices

[Chris Stubben]

Jens Oldeland wrote:
> 
> I am looking for a solution on how to sample increasing sizes of 
> "quadrats" in a list of matrices. Each matrix is a binary raster map and 
> I want to know if there is a 1 or only 0 s in the sampling unit 
> (quadrat, e.g. 4x4 cells). 
> 


You could just use a few loops to index the submatrices/quadrats, maybe
something like this.  

set.seed(7)
x<-matrix(sample(c(0,1), 256, replace=TRUE, prob=c(.8,.2)),  nrow=16,
byrow=TRUE)
# matrix size (16)
n<-nrow(x)
#quadrat sizes (1  2  4  8 16)
q<-  2^(0:sqrt(n))
# store results in list
quads<-vector("list", length(q) )
names(quads)<-q
for (i in 1:length(q) ) 
{
   z<-c()    
   for(j in 1:(n-q[i]+1) )    
   {
      for(k in 1:(n-q[i]+1))
      {
         y<-x[j:(j+q[i]-1), k:(k+q[i]-1)]
         #do something with submatrix
         z<-c(z, ifelse( sum(y)>0, 1,0))
         #z<-c(z, sum(y) ) 
        
      }
   }
   quads[[i]]<-z
}

#empty quads
sapply(quads, function(x) sum(x==0))
  1   2   4   8  16 
206  91   3   0   0 
# total number of quadrats sampled
sapply(quads, length)
#or
(n-q+1)^2


For random sampling, try replacing 1:(n-q[i]+1)  with  sample( 1:(n-q[i]+1),
replace=TRUE) ?

Also, if you have a list of matrices, then maybe combine the code above into
a function and then apply it to your list using lapply.


Chris Stubben

-- 
View this message in context: http://www.nabble.com/sampling-quadrats-of-increasing-size-in-a-list-of-matrices-tp24270434p24276950.html
Sent from the R help mailing list archive at Nabble.com.