evalEnse: Evaluate community assemblages predictions.
In dinilu/paleoCLMs-package: Miscelanea package for paleoCLMs project

Description Usage Arguments Details Value Author(s)

This function evaluate several measures of community assemblage for predictions. It take two community matrices (observed and predicted), a logical vector specifing test(training) datasets, and an option to specify if the data in the community matrices are presence/absence (binary=TRUE) or abundances (binary=FALSE).

1	evalEnse(obse, pred, tS, binary=TRUE)

`obse`	Matrix. Community matrix with observed values. Thay can be binary data (presence/absence) or abundance data.
`pred`	Matrix. Community matrix with predicted values. It should have the same size (and with the same column and row order) than obse.
`tS`	Logical vector. This logical vector should indicate which sites will be used as training (TRUE) and testing (FALSE).
`binary`	Logical value. This specify if the data are binary (TRUE) or abundance (FALSE).

This function doesn't check for consistency between data and binary parameter. If there is a mismatch between them the result won't be reliable.

Vector. The output is a small vector(matrix) with 6 elements: mean Jaccard index, standard deviation of Jaccard index, discrimination (as correlation) and calibration (as model efficiency) bewteen pair of sites dissimilarity, species richness discrimination (as correlation) and calibration (as model efficiency).

Model efficiency has been defined as the Nash-Sutcliffe Efficiency.

Diego Nieto Lugilde

dinilu/paleoCLMs-package documentation built on May 15, 2019, 8:46 a.m.