multifunc (version 0.9.4)

getRedundancy: getRedundancy

Description

getRedundancy examines which species have an effect on which function

Usage

getRedundancy(
  vars,
  species,
  data,
  negVars = NA,
  method = "lm",
  combine = "+",
  output = "effect",
  ...
)

Arguments

vars

Vector of column names of functions

species

Vector of column names of species

data

data frame with species presence/absence of values of functions

negVars

Vector of names of species for which a negative coefficient is actually a positive effect.

method

Fitting function for statistical models. Defaults to lm.

combine

How are species combined in the model? Defaults to "+" for additive combinations.

output

Will the output be sign of effect or "coefficient". Defaults to "effect"

...

Other arguments to be supplied to fitting function.

Value

Returns a matrix of functions and the effect of species on each. 1s, -1s, and 0s for "effect" or coefficients.

Details

getRedundancy takes a matrix of 1s,0s, and -1s, and depending on whether we're interested in positive, negative, or both types of interactions looks for the m-wise overlap between species and returns the overlap index for each combination. For species whose effect is not different from 0 at the alpha=0.05 level, a 0 is returned.

Examples

# NOT RUN {
data(all_biodepth)
allVars <- qw(biomassY3, root3, N.g.m2, light3, N.Soil, wood3, cotton3)

germany <- subset(all_biodepth, all_biodepth$location == "Germany")

vars <- whichVars(germany, allVars)
species <- relevantSp(germany, 26:ncol(germany))

# re-normalize N.Soil so that everything is on the same
# sign-scale (e.g. the maximum level of a function is
# the "best" function)
germany$N.Soil <- -1 * germany$N.Soil + max(germany$N.Soil, na.rm = TRUE)

res.list <- lapply(vars, function(x) sAICfun(x, species, germany))
names(res.list) <- vars

getRedundancy(vars, species, germany)
getRedundancy(vars, species, germany, output = "coef")



#########
# takes a vector of responses, the species that may cause them
# and returns a table of 1s, -1s, and 0s with regards to the kind of effect
# or a coefficient table, if asked for.  Arugments can take the form of the fitting function
# how variables are combined, and additional arguments to the fitting function
#########
# }