public void test2()
{
WriteLine("Generating the normal samples spacial points. ");
double[] mu1 = new double[] { 200, 0 }, mu2 = new double[] { -200, 0 };
double sd = 50;
int EachClusterSize = 100;
// SpacialPoint ExpectedCenter1 = new SpacialPoint(mu1);
// SpacialPoint ExpectedCenter2 = new SpacialPoint(mu2);
WriteLine($"2 hypothetical centers: [[{mu1[0]}, {mu1[1]}], [{mu2[0]},{mu2[1]}],");
int Counter = 10;
WriteLine("Average Normalized Deviations from centroid to centers");
while (--Counter != 0)
{
Point[] samples1 = SpatialPoint.NormalRandomPoints(mu1, sd, EachClusterSize);
Point[] samples2 = SpatialPoint.NormalRandomPoints(mu2, sd, EachClusterSize);
Point[] merged = new Point[samples1.Length + samples2.Length];
Array.Copy(samples1, 0, merged, 0, samples1.Length);
Array.Copy(samples2, 0, merged, samples1.Length, samples2.Length);
KMinSpanningTree Kmst = new KMinSpanningTree(merged);
WriteLine("------------------------Centroids:----------------------------------");
Point[][] SetOfClusters = MakeNestedArrayFomPointCollection(Kmst.KSpanningTree);
SaveTheClustersToJson($"tests_output/test2_clusters{Counter}.json", SetOfClusters);
}
}
public Task GetTestInstance()
{
TestRecord that = this;
Task r = new Task
(
() =>
{
Stopwatch sw = new Stopwatch();
SpatialPoint ExpectedCenter1 = new SpatialPoint(Mu1), ExpectedCenter2 = new SpatialPoint(Mu2);
Point[] samples1 = SpatialPoint.NormalRandomPoints(Mu1, 1, SizeofEachCluster);
Point[] samples2 = SpatialPoint.NormalRandomPoints(Mu2, 1, SizeofEachCluster);
Point[] merged = new Point[samples1.Length + samples2.Length];
Array.Copy(samples1, 0, merged, 0, samples1.Length);
Array.Copy(samples2, 0, merged, samples1.Length, samples2.Length);
sw.Start();
FullGraphClustering fg = new FullGraphClustering(merged);
Point c1 = fg.Centroid1;
Point c2 = fg.Centroid2;
double d1 = Math.Min(Point.Dis(c1, ExpectedCenter1), Point.Dis(c1, ExpectedCenter2));
double d2 = Math.Min(Point.Dis(c2, ExpectedCenter1), Point.Dis(c2, ExpectedCenter2));
double d = (d1 + d2) / 2;
d /= SDaway;
sw.Stop();
that.Register(d, sw.Elapsed.TotalMilliseconds);
}
);
return(r);
}
public void TestPointsClass()
{
Point p1 = new SpatialPoint(new int[] { 0, 0, 0 });
Point p2 = new SpatialPoint(new int[] { 1, 1, 1 });
Assert.IsTrue(Point.Dis(p1, p2) == Math.Sqrt(3));
}
public void TestClustering()
{
WriteLine("Generating the normal samples spacial points. ");
double[] mu1 = new double[] { 0, 0, 0 }, mu2 = new double[] { 200, 200, 200 };
double sd = 100;
int EachClusterSize = 500;
SpatialPoint ExpectedCenter1 = new SpatialPoint(mu1);
SpatialPoint ExpectedCenter2 = new SpatialPoint(mu2);
WriteLine($"2 hypothetical centers: [[{mu1[0]}, {mu1[1]}, {mu1[2]}], [{mu2[0]},{mu2[1]},{mu2[2]}]],");
int Counter = 100;
WriteLine("Average Normalized Deviations from centroid to centers");
double DeviationAverage = 0;
while (--Counter != 0)
{
Point[] samples1 = SpatialPoint.NormalRandomPoints(mu1, sd, EachClusterSize);
Point[] samples2 = SpatialPoint.NormalRandomPoints(mu2, sd, EachClusterSize);
Point[] merged = new Point[samples1.Length + samples2.Length];
Array.Copy(samples1, 0, merged, 0, samples1.Length);
Array.Copy(samples2, 0, merged, samples1.Length, samples2.Length);
FullGraphClustering fg = new FullGraphClustering(merged);
Point c1 = fg.Centroid1;
Point c2 = fg.Centroid2;
double d1 = Math.Min(Point.Dis(c1, ExpectedCenter1), Point.Dis(c1, ExpectedCenter2));
double d2 = Math.Min(Point.Dis(c2, ExpectedCenter1), Point.Dis(c2, ExpectedCenter2));
double d = (
d1 / (Math.Sqrt(mu1.Length) * sd)
+
d2 / (Math.Sqrt(mu2.Length) * sd)
) / 2;
WriteLine(
d
);
DeviationAverage += d / 100;
}
WriteLine($"Average Deviation: {DeviationAverage}");
}
