plot_bf {bite}R Documentation

Plots summary of Bayes Factors calculations

Description

Lolipop plot representing values of BF (Bayes Factor) scores for different models

Usage

plot_bf(m.liks, thr = 2, dir = c("vertical", "horizontal"),
  col = c("#d32f23", "#2e86ab", "#000000"), col.thr = c("#a6e1fa"),
  ax.lab = "log(BF)", main = "", rank = TRUE, dec = TRUE,
  group.pattern = NULL, cex = c(1.2, 1, 1), mod.lab = NULL,
  srt.lab = 0, adj.lab = 0, space = c(1.2, 0.8))

Arguments

m.liks

matrix of marginal likelihoods (see details)

thr

value of BF threshold for model selection (default is 2), Several thresholds can be given in the form of a vector

dir

string giving the direction of the plot ("vertical", "horizontal")

col

color of the lines and dots. Could be of size one or more. The first element applies to the best model, the second to the ones below thr[1] and so on...

col.thr

color for the threshold area (must be at the sime length as thr)

ax.lab

label of the axis.

main

an overall title for the plot

rank

logical: should models be ranked by BF scores? (default is TRUE)

dec

logical: should models be displayed in decreasing order? (default is TRUE)

group.pattern

Regular expression given a pattern to be matched in names(m.liks). The values matching that pattern will be grouped in the plot

cex

size of the dots and width of the lines. Could be of size one or more. The first element applies to the best model, the second to the ones below thr[1] and so on...

mod.lab

name of the models. If mod.lab = NULL the names of m.liks are used

srt.lab, adj.lab

rotation and justification of model names, see srt and adj in par

space

only evaluated if group.pattern != NULL, numeric(2) indicating the space between groups and the space between members of the same groups

Author(s)

Theo Gaboriau

Examples

## Fake marginal likelihood data
m.liks <- c(20 ,33, 56, 51, 55, 12)
names(m.liks) <- c("MBM-VBM", "MBM-VWN", "MBM-VOU", "MOU-VBM", "MOU-VWN", "MOU-VOU")

#Does not go well with default margin sizes
oldmar <- par()$mar
par(mar = c(5,1,4,6))
plot_bf(m.liks)
plot_bf(m.liks, thr = c(2,6), col.thr = c("#a2c5ac", "#ade1e5"),
 col = c("#d32f23", "#468189","#2e86ab","#000000"), cex = c(1.2,1,0.8,0.8))
plot_bf(m.liks, group.pattern = "MBM", rank = FALSE)
plot_bf(m.liks, group.pattern = "MOU", rank = TRUE)
plot_bf(m.liks, group.pattern = c("VWN", "VOU", "VBM"), rank = TRUE)
par(mar = c(6,5,1,2))
plot_bf(m.liks, dir = "horizontal", srt.lab = -60, adj.lab = c(0,0.8))
par(mar = oldmar)

[Package bite version 0.3 Index]