public bool Equals(KMeansPoint other)
{
if (ReferenceEquals(null, other))
{
return(false);
}
if (ReferenceEquals(this, other))
{
return(true);
}
return(other.X.Equals(X) && other.Y.Equals(Y));
}
public static IEnumerable <KMeansCluster> FindMatrixClusters(int matX, int matY, int clusterCount)
{
//Generating inputs
var input = new List <KMeansPoint>();
for (int i = 0; i < matX; i++)
{
for (int j = 0; j < matY; j++)
{
input.Add(new KMeansPoint(i, j));
}
}
//Generating clusters
var clusters = new List <KMeansCluster>();
for (int i = 0; i < clusterCount; i++)
{
var location = new KMeansPoint(Ran.Next(matX), Ran.Next(matY));
while (clusters.Any(item => location.Equals(item)))
{
location = new KMeansPoint(Ran.Next(matX), Ran.Next(matY));
}
clusters.Add(new KMeansCluster {
Location = location
});
}
bool running = true;
while (running)
{
//Assing inputs to clusters
foreach (KMeansCluster cluster in clusters)
{
cluster.AssignedInputs.Clear();
}
foreach (KMeansPoint point in input)
{
double minDistance = double.MaxValue;
KMeansCluster minCluster = null;
foreach (KMeansCluster cluster in clusters)
{
double distance = cluster.Location.CalculateDistance(point);
if (distance < minDistance)
{
minDistance = distance;
minCluster = cluster;
}
}
if (minCluster != null)
{
minCluster.AssignedInputs.Add(point);
}
}
running = false;
//Recalculate cluster positions
foreach (KMeansCluster cluster in clusters)
{
double newX = 0;
double newY = 0;
foreach (KMeansPoint point in cluster.AssignedInputs)
{
newX += point.X;
newY += point.Y;
}
newX = newX / cluster.AssignedInputs.Count;
newY = newY / cluster.AssignedInputs.Count;
if (Math.Abs(cluster.Location.X - newX) > 0.001 || Math.Abs(cluster.Location.Y - newY) > 0.001)
{
running = true;
}
cluster.Location.X = newX;
cluster.Location.Y = newY;
}
//Fixing zero clusters
if (running == false)
{
foreach (var cluster in clusters)
{
if (cluster.AssignedInputs.Count == 0)
{
var location = new KMeansPoint(Ran.Next(matX), Ran.Next(matY));
while (clusters.Any(item => location.Equals(item)))
{
location = new KMeansPoint(Ran.Next(matX), Ran.Next(matY));
}
cluster.Location = location;
}
}
}
}
return(clusters);
}
public double CalculateDistance(KMeansPoint point)
{
return(Math.Sqrt(Math.Pow(Math.Abs(X - point.X), 2) + Math.Pow(Math.Abs(Y - point.Y), 2)));
}