error.quantisation.somnum: Quantisation error for a Self-Organising Map fitted on vector...

In yasomi: Yet Another Self Organising Map Implementation

Description

Compute the quantisation error made by a fitted Self-Organising Map: this is the mean of the Euclidean distance between each observation and the prototype of its best matching unit.

Usage

## S3 method for class 'somnum'
error.quantisation(som, newdata, ...)

Arguments

som	an object of class "somnum"
newdata	an optional matrix or a data frame of observations compatible with the fitted som
...	not used

Value

If newdata is not given, the function returns the quantisation error made by the fitted som on the data used to fit it. When newdata is specified, the function returns the quantisation error of the fitted som on the corresponding data.

Author(s)

Fabrice Rossi

See Also

error.kaskilagus.somnum, som.tunecontrol

Examples

data(iris)
# scaling
data <- scale(iris[1:4])

# a small hexagonal grid
sg <- somgrid(xdim=7,ydim=7,topo="hex")

# fit the SOM
som <- batchsom(data,sg)

print(error.quantisation(som))

# a larger grid should in general give a lower quantisation error
sg <- somgrid(xdim=12,ydim=12,topo="hex")
som <- batchsom(data,sg)
print(error.quantisation(som))

yasomi documentation built on May 2, 2019, 5:59 p.m.