This script was written to examine the effect of combining botanical field surveys from multiple indepentent observers. The rational is that false-positive observations can be reduced by combining the species lists created by one or more observer. However, this increases the number of false-negative observations. The script bootstraps the selection of surveys and calculates a mean number of false-positive ands false-negative observations when the surveys are combined.
It takes an input matrix in which the columns are species and each row is the results of a survey. Where a species was not found in the survey the value in the column is 0. A 1 signifies that the species was correctly observered in the survey. A 2 signifies that the species was observed in the survey, but was, in fact, absent.
####usage
groupedSurveys.pl surveysFile outputFileName #replicates #surveys
The survey file is the input matrix.
replicates is the number of times to repeat the bootstraping.
surveys is the number of surveys in the input to combine.
The output file contains the following columns
- the replicate number
- the number of species found, assuming each combined survey has to have found that species
- number of true-positives in combined surveys
- the total number of species if just two surveys are combined
- number of false-positives in combined surveys
- the total number of false-positives if just two surveys are combined
- the total number of true-positives if just two surveys are combined
- accumulated sum of true-positives found
- accumulated sum of false-positives found
The file OldRiverBedInput.txt has been added as an example of an input file