rf.partial.prob: Random Forest probability scaled partial dependency plots

In rfUtilities: Random Forests Model Selection and Performance Evaluation

Description

Produces partial dependency plots with probability distribution based on scaled margin distances.

Usage

rf.partial.prob(x, pred.data, xname, which.class, w, prob = TRUE,
  plot = TRUE, smooth, conf = TRUE, smooth.parm = NULL,
  pts = FALSE, raw.line = FALSE, rug = FALSE, n.pt, xlab, ylab, main,
  ...)

Arguments

x	Object of class randomForest
pred.data	Training data.frame used for constructing the plot,
xname	Name of the variable for calculating partial dependence
which.class	The class to focus on
w	Weights to be used in averaging (if not supplied, mean is not weighted)
prob	Scale distances to probabilities
plot	(TRUE/FALSE) Plot results
smooth	c(spline, loess) Apply spline.smooth or loess to
conf	(TRUE/FALSE) Should confidence intervals be calculated for smoothing
smooth.parm	An appropriate smoothing parameter passed to loess or smooth.spline
pts	(FALSE/TRUE) Add raw points
raw.line	(FALSE/TRUE) Plot raw line (non-smoothed)
rug	Draw hash marks on plot representing deciles of x
n.pt	Number of points on the grid for evaluating partial dependence.
xlab	x-axis plot label
ylab	y-axis plot label
main	Plot label for main
...	Additional graphical parameters passed to plot

Value

A list class object with fit x,y. If smooth=c("spline","loess") y represents smoothed scaled margin distance values

Author(s)

Jeffrey S. Evans <jeffrey_evans<at>tnc.org>

References

Evans J.S., M.A. Murphy, Z.A. Holden, S.A. Cushman (2011). Modeling species distribution and change using Random Forests CH.8 in Predictive Modeling in Landscape Ecology eds Drew, CA, Huettmann F, Wiersma Y. Springer

Baruch-Mordo, S., J.S. Evans, J. Severson, J. D. Naugle, J. Kiesecker, J. Maestas, & M.J. Falkowski (2013) Saving sage-grouse from the trees: A proactive solution to reducing a key threat to a candidate species Biological Conservation 167:233-241

See Also

smooth.spline for smooth.spline details on spar smoothing argument

loess for loess details of span smoothing argument

Examples

 require(randomForest)
   data(iris)
   iris.rf <- randomForest(iris[,1:4], iris[,5])		
 	 
 # plot all parameters	 
 par(mfrow=c(2,2))
   for(i in names(iris)[1:4]) {     
     rf.partial.prob(iris.rf, iris, i, "setosa", smooth="spline", 
                     n.pt=70, smooth.parm = 0.5)
    }

 # Plot spline and loess smoothing for one parameter, with raw points and line
 par(mfrow=c(1,2))	 
   rf.partial.prob(x = iris.rf, pred.data = iris, xname = "Sepal.Length", 
                   which.class = "setosa", smooth = "spline", smooth.parm = 0.5,
   				  n.pt = 70, pts = TRUE, raw.line = TRUE, rug = TRUE)
   				
   rf.partial.prob(x = iris.rf, pred.data = iris, xname = "Sepal.Length", 
                   which.class = "setosa", smooth = "loess", smooth.parm = 0.20,
   				  n.pt = 70, pts = TRUE, raw.line = TRUE, rug = TRUE)

rfUtilities documentation built on Oct. 3, 2019, 9:04 a.m.