ImpPlot {statVisual}

Plot of Variable Importance

Description

Plot of variable importance based on results from randomForest or gbm.

Usage

ImpPlot(model, 
	theme_classic = TRUE, 
	n.trees = NULL, 
	addThemeFlag = TRUE,
	...)

Arguments

model

An object returned by randomForest or gbm

theme_classic

logical. Use classic background without grids (default: TRUE).

n.trees

integer. The number of trees used to generate the plot used in the function summary.gbm in the R library gbm. Only the first n.trees trees will be used.

addThemeFlag

logical. Indicates if light blue background and white grid should be added to the figure.

...

other input parameters for facet & theme

Value

A list with 9 elements. data, layers, scales, mapping, theme, coordinates, facet plot_env, and labels.

Author(s)

Wenfei Zhang <[e-mail]>, Weiliang Qiu <[e-mail]>, Xuan Lin <[e-mail]>, Donghui Zhang <[e-mail]>

Examples

library(dplyr)
library(randomForest)
library(tibble)


data(esSim)
print(esSim)

# expression data
dat = exprs(esSim)
print(dim(dat))
print(dat[1:2,])

# phenotype data
pDat = pData(esSim)
print(dim(pDat))
print(pDat[1:2,])

# feature data
fDat = fData(esSim)
print(dim(fDat))
print(fDat[1:2,])

# choose the first 6 probes (3 OE probes, 2 UE probes, and 1 NE probe)
pDat$probe1 = dat[1,]
pDat$probe2 = dat[2,]
pDat$probe3 = dat[3,]
pDat$probe4 = dat[4,]
pDat$probe5 = dat[5,]
pDat$probe6 = dat[6,]

print(pDat[1:2, ])

# check histograms of probe 1 expression in cases and controls
print(table(pDat$grp, useNA = "ifany"))

pDat$grp = factor(pDat$grp)


rf_m = randomForest(
  x = pDat[, c(3:8)], 
  y = pDat$grp, 
  importance = TRUE, proximity = TRUE
)


statVisual(type = 'ImpPlot', model = rf_m)

ImpPlot(model = rf_m)