smaa (version 0.3-3)

smaa: One-stage SMAA analysis

Description

Calculate SMAA decision indices based on a set of samples from the criteria values distribution and a set of samples from the feasible weight space.

Usage

smaa(meas, pref)

Value

ra

Rank acceptabilities (see smaa.ra).

cw

Central weights (see smaa.cw).

Arguments

meas

Criteria measurements. An \(N \times m \times n\) array, where meas[i,,] is a matrix where the m alternatives are the rows and the n criteria the columns. The values must be standardized measurements (i.e. after application of the partial value function). smaa.pvf provides a convenience method to standardize partial values.

pref

Weights. An \(N \times n\) array, where pref[i,] is a normalized weight vector.

Author

Gert van Valkenhoef

Details

The one-stage method does not store the alternatives' values or the raw rankings. Instead, only standard summary metrics are provided.

See Also

smaa.pvf

Examples

N <- 1E4; m <- 2; n <- 3
meas <- dget(system.file("extdata/thrombo-meas.txt.gz", package="smaa"))

# Read weights from file
pref <- dget(system.file("extdata/thrombo-weights-nopref.txt.gz", package="smaa"))
# Alternatively, sample weights using hitandrun:
# library(hitandrun)
# pref <- simplex.sample(n, N)$samples

# Calculate SMAA metrics
result <- smaa(meas, pref)
print(result)
plot(result)

result <- smaa(meas, c(0.5, 0.2, 0.3))
print(result)