Visual Analysis for Cluster Tendency Assessment (VAT/iVAT)

Implements Visual Analysis for Cluster Tendency Assessment (VAT; Bezdek and Hathaway, 2002) and Improved Visual Analysis for Cluster Tendency Assessment (iVAT; Wang et al, 2010).

manip, cluster
VAT(x, ...)
iVAT(x, ...)
a dist object.
further arguments are passed on to pimage.

path_dist redefines the distance between two objects as the minimum over the largest distances in all possible paths between the objects as used for iVAT.


Bezdek, J.C. and Hathaway, R.J. (2002): VAT: a tool for visual assessment of (cluster) tendency. Proceedings of the 2002 International Joint Conference on Neural Networks (IJCNN '02), Volume: 3, 2225--2230.

Havens, T.C. and Bezdek, J.C. (2012): An Efficient Formulation of the Improved Visual Assessment of Cluster Tendency (iVAT) Algorithm, IEEE Transactions on Knowledge and Data Engineering, 24(5), 813--822.

Wang L., U.T.V. Nguyen, J.C. Bezdek, C.A. Leckie and K. Ramamohanarao (2010): iVAT and aVAT: Enhanced Visual Analysis for Cluster Tendency Assessment, Proceedings of the PAKDD 2010, Part I, LNAI 6118, 16--27.

See Also

seriate, pimage, create_lines_data.

  • path_dist
  • iVAT
  • VAT
## lines data set from Havens and Bezdek (2011) 
x <- create_lines_data(250)
plot(x, xlim=c(-5,5), ylim=c(-3,3), cex=.2)  
d <- dist(x)

## create regular VAT
VAT(d, colorkey = TRUE, main = "VAT")
## same as: pimage(d, seriate(d, "VAT"))  

## create iVAT which shows visually the three lines
iVAT(d, main = "iVAT")
## same as:
## d_path <- path_dist(d)
## pimage(d_path, seriate(d_path, "VAT"))  

## compare with dissplot (shows banded structures and relationship between 
## center line and the two outer lines)
dissplot(d, method="OLO_single", main = "Dissplot", col = bluered(100, bias = .5))

## compare with optimally reordered heatmap
hmap(d, method="OLO_single", main = "Heat map (opt. leaf ordering)", 
col = bluered(100, bias = .5))
Documentation reproduced from package seriation, version 1.2-2, License: GPL-3

Community examples

Looks like there are no examples yet.