Package Safi': R Topics Documented
Package Safi': R Topics Documented
R topics documented:
safi-package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
accessSafiDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
createSafiDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
demo3steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
plot.safidesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
plot.safimodel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
safiModel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
splitSafiDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Index 9
1
2 accessSafiDesign
safi-package safi
Description
Author(s)
References
Fruth, J.; Roustant, O.; Kuhnt, S. (2014) Sequential designs for sensitivity analysis of functional in-
puts in computer experiments, Reliability Engineering & System Safety, doi: 10.1016/j.ress.2014.07.018,
preprint on HAL: http://hal.archives-ouvertes.fr/hal-00943509.
Description
The functional input design is printed at a chosen number of time points to allow for data transfer
to the output model.
Usage
accessSafiDesign(s.d, n.timepoints)
Arguments
s.d safidesign-object
n.timepoints number of desired time points at which to evaluate the functions, either one value
(the same number for all inputs) or a vector of length d.f (different numbers for
the inputs)
Value
Examples
# 1 functional input
s.d <- createSafiDesign(d.f = 1)
s.d2 <- splitSafiDesign(s.d = s.d, new.split.points = list(c(0.25, 0.75)))
accessSafiDesign(s.d2, n.timepoints = 20)
# 3 functional inputs
s.d <- createSafiDesign(d.f = 3)
s.d2 <- splitSafiDesign(s.d = s.d, new.split.points = list(c(0.5), c(0.25, 0.75),
c(0.25, 0.5, 0.75)))
accessSafiDesign(s.d = s.d2, n.timepoints = c(2, 4, 10))
Description
An initial design object is created.
Usage
createSafiDesign(method = "SB", mirrored.runs.included = FALSE, d.f = 1, variable.names)
Arguments
method a character string specifying the evaluation method to be chose from "SB" and
"other". Currently only "SB", sequential bifurcation, is implemented. For
"other" the DoE has to be added to the object manually.
mirrored.runs.included
boolean. If TRUE mirror runs will be included in the design.
d.f number of (functional) inputs
variable.names optional variable names (if NULL names are set to x1,x2,...)
Value
safidesign object containing DoE, split.points and d.f, mirrored.runs.included and method
References
Fruth, J.; Roustant, O.; Kuhnt, S. (2014) Sequential designs for sensitivity analysis of functional in-
puts in computer experiments, Reliability Engineering & System Safety, doi: 10.1016/j.ress.2014.07.018,
preprint on HAL: http://hal.archives-ouvertes.fr/hal-00943509.
Bettonvil, B. (1995) Factor screening by sequential bifurcation, Communications in Statistics-
Simulation and Computation, 24, 165-185.
4 demo3steps
Examples
# one input
s.d <- createSafiDesign()
s.d
# two inputs
s.d <- createSafiDesign(d.f = 2)
s.d
Description
Demo of a complete sequential sensitivity analysis with 2 functional and 1 scalar input in three
sequential steps.
Details
The procedure is performed on an artificial underlying function with two functional inputs g1 and
g2 and one scalar input x.
Z 1/3 Z 1
f (g1, g2, x) = (3 − 9t)g1(t)dt − [1/30g2(t) + 3]3 dt + 8/10 sin(x)
0 3/10
Description
Plotting the functional inputs against the functional domain
Usage
## S3 method for class 'safidesign'
plot(x, runs = NULL, ...)
Arguments
x safidesign-object
runs vector of runs that shall be plotted, if not provided all runs are plotted
... optional graphical parameters
Value
One plot for each input, containing all design functions given in runs.
Examples
# 1 functional input
s.d <- createSafiDesign(d.f = 1)
plot(s.d)
# 3 functional inputs
s.d <- createSafiDesign(d.f = 3)
s.d2 <- splitSafiDesign(s.d = s.d, new.split.points = list(c(0.5), c(0.25, 0.75),
c(0.25, 0.5, 0.75)))
plot(s.d2)
plot(s.d2, runs = 1:2) # only first two runs
Description
Barplots of the functional sensitivities from a safimodel object.
Usage
## S3 method for class 'safimodel'
plot(x, ylim = NULL, ...)
6 safiModel
Arguments
x safimodel object
ylim limits for the y-axis. If NULL, limits are computed internally.
... optional graphical parameters
Value
One plot for each functional input.
References
Fruth, J.; Roustant, O.; Kuhnt, S. (2014) Sequential designs for sensitivity analysis of functional in-
puts in computer experiments, Reliability Engineering & System Safety, doi: 10.1016/j.ress.2014.07.018,
preprint on HAL: http://hal.archives-ouvertes.fr/hal-00943509.
Examples
### simple example
Description
computes normalized regression indices for the sensitivity analysis of functional inputs
safiModel 7
Usage
safiModel(s.d, y)
Arguments
s.d safidesign-object
y model response
Details
If the design was created with method "SB" the coefficients are computed via sequential bifurcation,
for method "other" via least squares estimation.
Value
safimodel object containing the design and the computed coefficients
References
Fruth, J.; Roustant, O.; Kuhnt, S. (2014) Sequential designs for sensitivity analysis of functional in-
puts in computer experiments, Reliability Engineering & System Safety, doi: 10.1016/j.ress.2014.07.018,
preprint on HAL: http://hal.archives-ouvertes.fr/hal-00943509.
Examples
Description
The functional domains of the input functions are split up at given points and a corresponding DoE
is generated.
Usage
splitSafiDesign(s.d, new.split.points)
Arguments
s.d safidesign object
new.split.points
list of new split points
Details
For method "other" no runs are added.
Value
safidesign-object
References
Fruth, J.; Roustant, O.; Kuhnt, S. (2014) Sequential designs for sensitivity analysis of functional in-
puts in computer experiments, Reliability Engineering & System Safety, doi: 10.1016/j.ress.2014.07.018,
preprint on HAL: http://hal.archives-ouvertes.fr/hal-00943509.
Bettonvil, B. (1995) Factor screening by sequential bifurcation, Communications in Statistics-
Simulation and Computation, 24, 165-185.
Examples
# two inputs
s.d <- createSafiDesign(d.f = 2)
s.d
s.d2 <- splitSafiDesign(s.d = s.d, new.split.points = list(c(0.5), c(0.25, 0.75)))
s.d2
∗Topic \textasciitildekwd1
createSafiDesign, 3
splitSafiDesign, 8
∗Topic \textasciitildekwd2
createSafiDesign, 3
splitSafiDesign, 8
∗Topic datasets
demo3steps, 4
∗Topic package
safi-package, 2
accessSafiDesign, 2
createSafiDesign, 3
demo3steps, 4
plot.safidesign, 5
plot.safimodel, 5
safi (safi-package), 2
safi-package, 2
safiModel, 6
splitSafiDesign, 8