Research Article
VIASCKDE Index: A Novel Internal Cluster Validity Index for Arbitrary-Shaped Clusters Based on the Kernel Density Estimation
| Input; X, labels | | Output; VIASCKDE | | KDE_X←Gaussian_KDE(X.T) | | for k = 1 to size (unique (labels)) do | | data_of_Cluster_k←X[labels = k] ►data belongs to cluster k | | data_not_belong_k_X[labels ≠ k] ►data not belongs to cluster k | | kde_k←MinMaxNormalization (KDE_X [labels = k]) ►KDE of each | | ►data of cluster k and normalize them | | for i = 1 to size (data_of_k)do | | ai ← closed_data (data_of_cluster_k) ►distance from ithdata | | ►to closed one in Cluster k | | bi ← closed_data (k_ait_olmagan_verileri) ►distance from ith | | ►data to closest one that does not in the cluster k | | CoSeDi = kde_k[i][(bi−ai)/max(ai, bi0)] ►Compactness and | | ►Separation of data i | | CoSeCk = mean (CoSeDk) ►Compactness-separation of Cluster k | | VIASCKDE = ►Overall VIASCKDE value of Clustering |
|