map.significance: Compute Significance of Features

In popsom: An Efficient Implementation of Kohonen's Self-Organizing Maps (SOMs) with Starburst Visualizations

Description

Computes the relative significance of each feature and plots it.

Usage

map.significance(map,graphics=TRUE,feature.labels=TRUE)

Arguments

map	An object of type 'map'.
graphics	A switch that controls whether a plot is generated or not.
feature.labels	A switch to allow the plotting of feature names vs feature indices.

Value

If graphics=FALSE a vector containing the significance for each feature is returned.

Note

We use a Bayesian approach to compute the relative significance of features based on variance.

Author(s)

Lutz Hamel

References

Bayesian Probability Approach to Feature Significance for Infrared Spectra of Bacteria, Lutz Hamel, Chris W. Brown, Applied Spectroscopy, Volume 66, Number 1, 2012.

Examples

data(iris)

df <- subset(iris,select=-Species)
labels <- subset(iris,select=Species)

m <- map.build(df,labels,xdim=15,ydim=10,train=10000)

## show the relative feature significance for each feature
df <- data.frame(names(df),map.significance(m,graphics=FALSE))
names(df) <- c("Features","Significance")
df

## display the relative feature significance graphically
map.significance(m)

popsom documentation built on Dec. 21, 2021, 1:07 a.m.