[R-sig-eco] Randomizing matrices

[Peter Solymos]

Allan,

Simply defining the dimension might work:
dim(v) <- dim(trial2)
but it is not clear what you are trying to achieve with the rep(...,
1000) part. It won't permute the matrix 1000 times but repeat same
values. You might want to have a look at oecosimu in vegan which
calculates the distribution given you SSD statistics, and you can use
many different matrix permutation algorithms although it seems to me
that you are just shuffling the cells.

Cheers,

Peter

--
Péter Sólymos, Dept Biol Sci, Univ Alberta, T6G 2E9, Canada AB
solymos at ualberta.ca, Ph 780.492.8534, http://psolymos.github.com
Alberta Biodiversity Monitoring Institute, http://www.abmi.ca
Boreal Avian Modelling Project, http://www.borealbirds.ca


On Sat, Aug 18, 2012 at 7:05 AM, Allan Edelsparre <aedelspa at uoguelph.ca> wrote:
> Dear R ecologists,
>
> I'm trying to figure out a way to calculate the sum of squared
> distances (SSD) between two matrices, where one matrix is held
> constant and the other is randomized. So far I have been able to get
> the syntax together to obtain my observed SSD, but the problem for me
> is obtain SSD from 1000 randomizations to obtain a distribution. Below
> is my syntax, all data were standardized a priori, each matrix is
> 6columns x 363rows.
>
> For now the issue seems to be that the 'sample' command destroys the
> matrix and turns it into a vector. I've tried to remedy this but now
> it seems I may have one huge matrix from my randomizations rather than
> 1000 of the original size.
>
> data1<-read.table("StandDataTime1and2.txt", header=TRUE)
> data1<-data.frame(data1)
>
> #split data into two matrices 6 X 363
> #####trial1
> trial1<-cbind(data1$fe1, data1$fr1, data1$ar1, data1$mid1, data$bot1,
> data$surf1)
> trial1<-as.matrix(trial1)
>
> ###trial2
> trial2<-cbind(data1$fe2, data1$fr2, data1$ar2, data1$mid2, data1$bot2,
> standardizedData$surf2)
> trial2<-as.matrix(trial2)
>
> ##calculate euclidean distance between matrices
> library(fields)
> distances<-rdist(trial1, trial2)
>
> ###getting the observed value SSD
> squared<-distances %*% distances
> ssd<-sum(squared)
> ssd
> ###observed value is 14301499
>
> ###########################################################################################
> Below is where the problem is occuring
>
>
> ########permuting one block (trial2) and calculating a 95%
> distribution from 1000 randomizations
> v <- rep(sample(trial2 <- as.matrix(trial2)),1000)
> ## Now run those thousand values through the equation to calculate the ssd2
>
> for (i in v) {
>   dist2 <-rdist(trial1,[i])
>   squared2<-dist2^2
>   ssd2<-sum(squared2)
>   print(ssd2)
> }
>
> Working with matrices seems to add to the level of difficulty. Any
> help here would be greatly appreciated. Thanks in advance for any
> help.
>
> Allan
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>