hypervolume_general_model: Generates hypervolume by sampling from arbitrary model...

In dc165/Hypervolume-Dev: High Dimensional Geometry and Set Operations Using Kernel Density Estimation, Support Vector Machines, and Convex Hulls

Description

Uses rejection sampling to obtain predicted values of a model object at uniformly random points within a range box, then converts output to a hypervolume.

Usage

hypervolume_general_model(model, name = NULL, verbose = TRUE, 
  data = NULL, range.box = NULL, num.samples = ceiling(10^(3 + sqrt(ncol(range.box)))), 
  chunk.size = 10000, min.value = 0, ...)

Arguments

model	Any model object which can be used within a predict(model, newdata, ...) call.
name	Name of the output hypervolume
verbose	If TRUE, prints diagnostic output.
data	If not NULL, used to specify range.box=padded_range(data).
range.box	A 2 x n matrix, where n is the number of dimensions of the data, and the first row corresponds to a lower limit and the second row to an upper limit. Each column is thus the low and high limits of the range box along each axis. Can be generated via padded_range.
num.samples	Number of samples to draw from the range box.
chunk.size	Number of samples to process in each predict call. Changing this value may affect the speed of function return but not the returned values.
min.value	If TRUE, discards sampled values below this threshold. Effectively used to set hypervolume boundaries.
...	Other arguments to be passed to predict, e.g. type='response'.

Value

A Hypervolume-class object corresponding to retained values within the hyperbox of interest.

Examples

data(iris)
iris[,"Species"] <- iris[,"Species"] == "setosa"
m_glm = glm(Species~.,data=iris)

hv_general_glm = hypervolume_general_model(m_glm, 
  range.box=padded_range(iris[,1:4]),type='response')
plot(hv_general_glm)	
	

dc165/Hypervolume-Dev documentation built on Dec. 13, 2020, 6:02 p.m.