bivariate.partialDependence: Bivariate partial-dependency plot

In rfUtilities: Random Forests Model Selection and Performance Evaluation

Description

Bivariate partial dependence provides a graphical depiction of the marginal effect of two variables on the class probability (classification) or response (regression)

Usage

bivariate.partialDependence(x, pred.data, v1, v2, grid.size = 20,
  which.class = 2, plot = TRUE, col.ramp = c("#ffffff", "#2a2a2a"),
  ncols = 20, ...)

Arguments

x	random forest object
pred.data	data.frame of independent variables used in model
v1	Variable 1 used in partial dependency
v2	Variable 2 used in partial dependency
grid.size	Number of grid cells (NxN) to integrate partial dependency for
which.class	Index of class probability (only if classification)
plot	(TRUE/FALSE) Plot 3D surface
col.ramp	Colors used in building color ramp
ncols	Number of colors in color ramp
...	Arguments passed to persp

Value

A list object with vectors of v1 (p1) and v2 (p2) and a matrix (estimate), estimate of the averaged estimates.

Note

In deriving the partial-dependence, at each plotted point, the background variables are held at their median values

Author(s)

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

References

Friedman, J.H. (2001) Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics 19(1)

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

persp for persp ... plotting options

Examples

 library(randomForest)
   data(iris)
   iris$Species <- ifelse( iris$Species == "versicolor", 1, 0 ) 
    
 # Add some noise
 idx1 <- which(iris$Species %in% 1)
 idx0 <- which( iris$Species %in% 0)
   iris$Species[sample(idx1, 2)] <- 0
   iris$Species[sample(idx0, 2)] <- 1
  
 # Specify model  
 y = iris[,"Species"] 
 x = iris[,1:4]
 
 set.seed(4364)  
 ( rf.mdl1 <- randomForest(x=x, y=factor(y)) )
 
 
 ( bvpd <- bivariate.partialDependence(rf.mdl1, iris, 
                   v1 = "Petal.Length", v2 = "Petal.Width", shade = 0.6,
                   grid.size = 20, ncols=100, border=NA, col.ramp=c("green","blue") ) ) 
	

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