Using Clustering Methods for Identifying Blood Donors Behavior

This study uses clustering method to describe blood donor behaviors. The goal of this clustering is provide the capability to determine blood donor segmentation for increasing the accuracy of manager predictions. We use k-means clustering algorithm for number of cluster 2, 3, 4, 5 and 6 and then we use the Dunn's index for distinguishing the optimal number of cluster. The results of Dunn's index show that optimal number of cluster is 2 which two-step clustering algorithm confirms it. The results of blood donor clustering describe that men have a tendency to donate their blood

کلیدواژه ها:

Blood Donor Behavior ، Clustering ، K-means Method ، Two-step Method