analysis.sensitivity {multisensi}

Runs a series of sensitivity analyses by a function from the sensitivity package

Description

The analysis.sensitivity function runs a series of sensitivity analyses on the columns of a data.frame, using a method implemented in the sensitivity package.

Usage

analysis.sensitivity(Y, plan, nbcomp = 2, sigma.car = NULL,
                     analysis.args = list(keep.outputs = FALSE))

Arguments

Y

a data.frame of output variables or principal components.

plan

an object containing the design. It must be created by a function from the sensitivity package with argument model=NULL.

nbcomp

the number of Y variables to analyse (the first nbcomp variables of Y will be analysed).

sigma.car

NULL or sum of squares of Y. If not NULL, compute the Generalised Sensitivity Indices (saved in the last column of the data.frame mSI/tSI/iSI outputs.

analysis.args

a list of arguments. If it contains keep.outputs=TRUE, the outputs associated with the analysis of each variable are returned (see section Value).

Details

The argument plan must be an object created by a method implemented in the sensitivity package. Thus it belongs to a class such as morris or fast99. The name of the class is stored in the element call.info$fct of the output returned by analysis.sensitivity.

Value

A list containing:

SI

data.frame of sensitivity indices or other importance measures returned by the function from the sensitivity package used. Sometimes empty but kept for compatibility reasons.

mSI

data.frame of first-order sensitivity indices

tSI

data.frame of total sensitivity indices

iSI

data.frame of interaction sensitivity indices

inertia

empty (kept for compatibility reasons)

indic.fact

0-1 matrix to indicate the factors associated with each factorial effect

Hpredict

empty (kept for compatibility reasons)

outputkept

if analysis.args$keep.outputs=TRUE, list of the outputs returned by the sensitivity analysis performed on each variable

call.info

list with first element analysis="sensitivity" and second element fct storing the class name of the argument plan

Examples

# Test case : the Winter Wheat Dynamic Models (WWDM)
library(sensitivity) # to use fast99
# input factors design
data(biomasseX)
# input climate variable
data(Climat)

# example of the sensitivity:fast99 function
# design
newplan <- fast99(model = NULL, factors = names(biomasseX), n = 100, 
            q = "qunif", q.arg = list(list(min = 0.9, max = 2.8),
                         list(min = 0.9, max = 0.99),
                         list(min = 0.6, max = 0.8),
                         list(min = 3, max = 12),
                         list(min = 0.0035, max = 0.01),
                         list(min = 0.0011, max = 0.0025),
                         list(min = 700, max = 1100)))
# simulations
wwdm.Y <- simulmodel(model=biomasse, plan=newplan$X, climdata=Climat)
# analysis
res <- analysis.sensitivity(data.frame(wwdm.Y), plan=newplan, nbcomp=4)