Overall Desirability

Description

This is a function to calculate the desirability for each response as well as the overall desirability.
The resulting data.frame can be used to plot the overall as well as the desirabilities for each response.
This function serves for a visualization of the desirability approach for multiple response optimization.

Usage

1
overall(fdo, steps = 20, constraints, ...)

Arguments

fdo

needs to be an object of class facDesign containing fits and desires

steps

numeric value - points per factor to be evaluated –> specifies also the grid size.

constraints

list - constraints for the factors in coded values such as list(A > 0.5, B < 0.2).

...

further arguments.

Value

overall returns a data.frame with a column for each factor, desirability for each response and a column for the overall desirability.

Author(s)

Thomas Roth thomas.roth@tu-berlin.de

References

see desirability.

See Also

facDesign
rsmDesign
desirability
http://www.r-qualitytools.org/html/Improve.html

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
#arbitrary example with random data!!!
rsdo = rsmDesign(k = 2, blocks = 2, alpha = "both")
set.seed(123)
response(rsdo) = data.frame(y = rnorm(nrow(rsdo)), y2 = rnorm(nrow(rsdo)))
fits(rsdo) = lm(y ~ A*B + I(A^2) + I(B^2), data = rsdo)
fits(rsdo) = lm(y2 ~ A*B + I(A^2) + I(B^2), data = rsdo)
desires(rsdo) = desirability(y, -1, 2, scale = c(1, 1), target = "max")
desires(rsdo) = desirability(y2, -1, 0, scale = c(1, 1), target = "min")
dVals = overall(rsdo, steps = 10, constraints = list(A = c(-0.5,1), 
                B = c(0, 1)))

##Uncomment for visualization of desirabilities
#require(lattice)
#contourplot(y ~ A*B,  data = dVals) #desirability of y
#contourplot(y2 ~ A*B,  data = dVals) #desirability of y2
#wireframe(overall ~ A*B, shade = TRUE, data = dVals)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.