spind (version 2.1.3)

rvi.plot: Relative Variable Importance

Description

Creates model selection tables, calculates and plots relative variable importance based on the scale level of a given model.

Usage

rvi.plot(formula, family, data, coord, maxlevel, detail = TRUE,
  wavelet = "haar", wtrafo = "dwt", n.eff = NULL, trace = FALSE,
  customize_plot = NULL)

Arguments

formula

A model formula

family

gaussian, binomial, and poisson are supported.

data

A data frame or set of vectors of equal length.

coord

X,Y coordinates for each observation. Coordinates should be consecutive integers.

maxlevel

An integer for maximum scale level

detail

Remove smooth wavelets? If TRUE, only detail components are analyzed. If set to FALSE, smooth and detail components are analyzed. Default is TRUE.

wavelet

Type of wavelet: haar, d4, or la8

wtrafo

Type of wavelet transform: dwt or modwt

n.eff

A numeric value of effective sample size

trace

Should R print progress updates to the console? Default is FALSE

customize_plot

Additional plotting parameters passed to ggplot

Value

A matrix containing the relative importance of each variable in the regression at each value of the scale level.

Details

Calculates the relative importance of each variable using multi-model inference methods in a wavelet multi-resolution regression framework implemented in mmiWMRR. The scale level dependent results are then graphically displayed.

Examples

# NOT RUN {
data(carlinadata)
coords<- carlinadata[,4:5]

# }
# NOT RUN {
wrm<- WRM(carlina.horrida ~ aridity + land.use,"poisson",
              carlinadata,coords,level=1,wavelet="d4")

mmi<- mmiWMRR(wrm,data=carlinadata,scale=3,detail=T)



# Plot scale-dependent relative variable importance
rvi.plot(carlina.horrida ~ aridity + land.use,"poisson",
         carlinadata,coords,maxlevel=4,detail=TRUE,wavelet="d4")
# }