R code error implementing co-ocurrence models

[drcala]

Hi!, I am trying to implement the co-ocurrence model for two binary datasets that I have. But when I am calling the "cooc_null_model" function, R gives me an error like this; "0%Error in speciesData[2, 1] : subscript out of bounds". Below I copied my code to illustrate the problem.

Thank you in advance

no_i<-read.csv("no invaded.csv", header = TRUE) inv<-read.csv("invaded.csv", header = TRUE)

media de la abundancia de cada sp en las parcelas no invadidas

row_weights_non<-rowMeans(no_i)

media de la abundancia de cada sp en las parcelas no invadidas

row_weights_inv<-rowMeans(inv)

col_weights<-rep(1, times=8)

myModel_inv<- cooc_null_model(speciesData="binary_in.csv", algo="sim10", suppressProg=TRUE, algoOpts = list(rowWeights=row_weights_inv,colWeights=col_weights))

[ngotelli]

Can you please send me one of the data sets that is throwing this error? Thanks.

[drcala]

binary_in.xlsx Thank you for your time

[ngotelli]

Looking quickly at your data set, I see several rows of zero, in which the species never occurs. That is probably what is throwing the error. Try deleting those rows and see if the analysis works.

[drcala]

The problem is that the file I sent you is the fragment of bigger one that has two types of plot in it. One with an invasive species as dominant and the other without this dominance. I am trying to analyse both types separately and see what results I have. But yes, I will do what you advice. There is no point of maintaining species that are not ocurring. Thank you

[drcala]

Hi again. I deleted the species that never occur and it is throwing the same error. It seems as I am asking to do something that is out of the boundaries of my data.

Thank you in advance

[emhart]

@drcala I'll take a look and see if I can track down what's throwing the error.

[drcala]

Ok, many thanks. Cheers

[emhart]

The error is actually due to how you're calling it. You're passing a string as the data matrix, not the actual object that you read in. You have to read the file "binary_in.csv" in to the matrix. It's giving you subscript out of bounds because it's expecting a dataframe, not a string (a string has no bounds). Here's an example based on the data you posted that worked for me....

testin <- read.csv("~/Documents/binary_in.csv",header=T,row.names = 1)
fixed_mat <- testin[apply(testin,1,sum) > 0,]

row_weights<-rowMeans(fixed_mat)
col_weights<-rep(1, times=6)

myModel_inv<- cooc_null_model(speciesData="fixed_mat", algo="sim10", suppressProg=TRUE,algoOpts = list(rowWeights=row_weights,colWeights=col_weights))

Here's my output:

> summary(myModel_inv)
Time Stamp:  Mon Nov  6 14:18:46 2017 
Reproducible:  FALSE 
Number of Replications:  1000 
Elapsed Time:  0.18 secs 
Metric:  c_score 
Algorithm:  sim10 
Observed Index:  0.98268 
Mean Of Simulated Index:  1.4734 
Variance Of Simulated Index:  0.045183 
Lower 95% (1-tail):  1.1225 
Upper 95% (1-tail):  1.8288 
Lower 95% (2-tail):  1.066 
Upper 95% (2-tail):  1.9118 
Lower-tail P =  0.014 
Upper-tail P =  0.986 
Observed metric > 14 simulated metrics 
Observed metric < 986 simulated metrics 
Observed metric = 0 simulated metrics 
Standardized Effect Size (SES):  -2.3088 

[ngotelli]

@emhart many thanks for quickly resolving this! @drcala I think you should be able to complete your analyses now, including the use of different indices with the EcoSim null models.

[drcala]

Oh Thank you @emhart !!! My bad! Now I see clearly what you mean. But what does the second line means? This one: fixed_mat <- testin[apply(testin,1,sum) > 0,] Many thanks

[drcala]

Oh, @emhart . I knew what you mean with the second line of the code. I already accounted for that issue because I deleted the rows (spp) that never occur. Thank you very much for your help