public void UnclusteredPoints()
{
var points = TestDataHelper.UniformRandomPoints(20000, 100, 1000000);
var ct = new ClusteringTendency(points, 5);
Assert.AreEqual(ClusteringTendency.ClusteringQuality.Unclustered, ct.HowClustered, $"Data should be unclustered, but was: {ct}");
}
public void WeaklyClustered()
{
// Mostly outliers with a few large clusters.
var clusteredPoints = TestData(Enumerable.Range(1, 3000).Select(i => i <= 5 ? 100 : 1).ToArray(), 100, 1000000);
var points = clusteredPoints.Points().ToList();
var ct = new ClusteringTendency(points, 5);
Assert.AreEqual(ClusteringTendency.ClusteringQuality.WeaklyClustered, ct.HowClustered, $"Data should be WeaklyClustered, but was: {ct}");
}
public void AllClustered()
{
// Fifty clusters all the same size.
var clusteredPoints = TestData(Enumerable.Range(1, 50).Select(i => 100).ToArray(), 100, 1000000);
var points = clusteredPoints.Points().ToList();
var ct = new ClusteringTendency(points, 5);
Assert.AreEqual(ClusteringTendency.ClusteringQuality.HighlyClustered, ct.HowClustered, $"Data should be HighlyClustered, but was: {ct}");
}
public void SingleOrMajorityCluster()
{
// Many clusters with one point and one with a whole lot.
var clusteredPoints = TestData(Enumerable.Range(1, 1000).Select(i => i == 1 ? 10000 : 1).ToArray(), 100, 1000000);
var points = clusteredPoints.Points().ToList();
var ct = new ClusteringTendency(points, 5);
var acceptable = ct.HowClustered == ClusteringTendency.ClusteringQuality.SinglyClustered ||
ct.HowClustered == ClusteringTendency.ClusteringQuality.MajorityClustered ||
ct.HowClustered == ClusteringTendency.ClusteringQuality.HighlyClustered;
Assert.IsTrue(acceptable, $"Data was: {ct}");
}
