public void test()
{
K_mean k = new K_mean();
k.doIt("data254.csv", "News", new List <Type>()
{
typeof(string)
}, 5, 400, "res123.csv");
}
static void Main(string[] args)
{
K_mean k = new K_mean();
k.doIt(@"../../Data/GExpress.txt", "News", new List <Type>()
{
typeof(string)
}, 5, 400, @"G_res.csv");
}
/// <summary>
/// after text preparation do the Kmean Process.
/// </summary>
/// <param name="data">double array of dataset</param>
/// <param name="k">num of clusters</param>
/// <param name="initial_centroids">initial centroids</param>
/// <param name="maxiter">Maximum iteration</param>
/// <param name="record_heterogeneity">heterogenity collection</param>
/// <returns>Centroids and Cluster assignmets</returns>
public Tuple <double[][], int[]> kmeans(double[][] data, int k, double[][] initial_centroids, int maxiter, ref List <double> record_heterogeneity)
{
K_mean km = new K_mean();
double[][] centroids = initial_centroids;
int[] prev_cluster_assignment = new int[] { -1 };
for (int i = 0; i < maxiter; i++)
{
#if verbose
if (i % 100 == 0)
{
Console.WriteLine("iter:\t" + i + " - " + (i / 100 + 1) * 100 + " \tof " + maxiter);
}
#endif
int[] cluster_assignment = assign_clusters(data, centroids);
centroids = revise_centroids(data, k, cluster_assignment);
if ((prev_cluster_assignment[0] != -1) && (Common.arraysEqual(cluster_assignment, prev_cluster_assignment)))
{
break;
}
#if verbose
if (prev_cluster_assignment[0] != -1)
{
int changed_clusters = Common.getArrayChanges(cluster_assignment, prev_cluster_assignment);
Console.WriteLine(changed_clusters + " elements changed their clusters.");
}
#endif
double score = compute_heterogeneity(data, k, centroids, cluster_assignment);
record_heterogeneity.Add(score);
prev_cluster_assignment = cluster_assignment;
}
#if verbose
for (int i = 0; i < record_heterogeneity.Count; i++)
{
Console.WriteLine("Heterogenity : " + record_heterogeneity[i]);
}
#endif
return(new Tuple <double[][], int[]>(centroids, prev_cluster_assignment));
}
