kmed (version 0.3.0)

barplotnum: Barplot of each cluster for numerical variables data set

Description

This function creates a barplot from a cluster result. A barplot indicates the location and dispersion of each cluster. The x-axis of the barplot is variable's mean, while the y-axis is the variable's name.

Usage

barplotnum(dataori, clust, nc = 1, alpha = 0.05)

Arguments

dataori

An original data set.

clust

A vector of cluster membership (see Details).

nc

A number of columns for the plot of all cluster (see Details).

alpha

A numeric number to set the significant level (between 0 and 0.2).

Value

Function returns a barplot.

Details

This is a marked barplot because some markers are added, i.e. a significant test, a population mean for each (numerical) variable. The significance test applies t-test between the population's mean and cluster's mean in every variable. The alpha is set in between 0 to 20%. If the population mean differs to the cluster's mean, the bar shade in the barplot also differs.

clust is a vector with the length equal to the number of objects (n), or the function will be an error otherwise. nc controls the layout (grid) of the plot. If nc = 1, the plot of each cluster is placed in a column. When the number of clusters is 6 and nc = 2, for example, the plot has a layout of 3-row and 2-column grids.

References

Leisch, F. (2008). Handbook of Data Visualization, Chapter Visualizing cluster analysis and finite mixture models, pp. 561-587. Springer Handbooks of Computational Statistics. Springer Verlag.

Dolnicar, S. and F. Leisch (2014). Using graphical statistics to better understand market segmentation solutions. International Journal of Market Research 56, 207-230.

Examples

# NOT RUN {
dat <- iris[,1:4]
memb <- cutree(hclust(dist(dat)),3)
barplotnum(dat, memb)
barplotnum(dat, memb, 2)

# }