Classification data

Cone-torus data

2 features
3 classes (cone, half-torus, Gaussian)
400 training and 400 testing data points
priors (0.25,0.25,0.5)

Download

a = 0.5

b = -p/3

Rotated checkerboard data

2 features
2 classes
a checkerboard tilted to b with side of the black squares equal to a, 0 < a < 1.

Download

Matlab code

Clustering data

Four-gauss data

2 true features +10 noise features
4 clusters
used in [Kuncheva & Hadjitodorov, 2004]

Download

Easy doughnut data

2 true features +10 noise features
2 clusters
used in [Kuncheva & Hadjitodorov, 2004]

Download

Difficult doughnut data

2 true features +10 noise features
2 clusters
used in [Kuncheva & Hadjitodorov, 2004]

Download

Boat data

2 features
3 clusters

Download

data

Noisy lines data

2 features
2 clusters

Download

data

Petals data

2 features
4 clusters

Download

data

Saturn data

2 features
2 clusters

Download

data

Regular data

2 features
16 clusters

Download

data

Half-ring data

2 features

2 clusters
From A. Topchy

Download

data

Spirals data

2 features

2 clusters
From A. Topchy

Download

data

Back

Latest modification to this page: Last modified 07/09/2010