COMPARISON OF FUZZY C-MEANS, FUZZY POSSIBILISTIC C-MEANS AND POSSIBILISTIC FUZZY C-MEANS ALGORITHMS ON THE DISTRIBUTION OF CONTRACEPTIVE USERS IN NTB PROVINCE

Abstract

Fuzzy Clustering is one of the parts from purposeful cluster method for group data by similarity characteristics. Advantages from method Fuzzy Clustering compared with method cluster other could make more detailed clusters. There are several methods in Fuzzy Clustering, including Fuzzy C-Means, Fuzzy Possibilistic C-Means and Possibilistic Fuzzy C-Means. Third method the could applied in various one field-field health that is for see scatter group user contraception. Contraception is tool for prevent proclaimed pregnancy for the success of the Family program Planning (KB) in push rate growth resident. Destination from study this is for compare third method Fuzzy Clustering with see score accuracy and see results cluster best formed based on score index validity Modified Partition Coefficient (MPC). Analysis result show that method Fuzzy C-Means is the best method seen from more MPC value height in each cluster. However, if seen from score iteration and time computation, method Fuzzy Possibilistic C-Means far more effective. There are 2 optimal clusters formed that is cluster 1 which describes spread user method contraception highest for tool contraception period length in NTB Province with amount districts reached 59. The dominating districts that is located on the island of Sumbawa. Whereas cluster 2 shows spread user tool contraception period tall, short that is It reaches 57 sub districts and is dominated by sub-districts on the island of Lombok.