IMPLEMENTASI DATA MINING DENGAN METODE CLUSTERING UNTUK MENGUKUR KECENDERUNGAN MEMILIH DAN TIDAK MEMILIH BAKAL CALON KEPALA DAERAH PADA PEMILIHAN KEPALA DAERAH

To manage the data elections are conducted every five years by direcelection of regional heads of both the local and regional levels of two particulaBengkulu city, from the results released by the General Elections Commissio(KPU) decreased from year to year. Voters needed to process the data that the method can be used to probe the hidden information of the Assets by the dataselector based on the polls (TPS) in every area of the city of Bengkulu, the mining will make the process of analyzing data to find hidden patterns or rules within thescope of the data set voters in this case using k-means algorithm (Unsupervised)is a non-hierarchical cluster methods, which attempt to partition the available data into two or more groups based on similarity and dissimilarity in particulawhere the same group is inserted into the same group and which hacharacteristics different to the other group by the center (mean point). In this tesby using Weka 3.7.5 to produce three groups as a parameter, which is a littvote, and many are based on the acquisition of each regional head candidate pepolling station (tps). From the results obtained by experiments conducted thefollowing results of votes which is 10% smaller, 20% of the vote was and 70% is lot of votes with Sum Squered Errors (SSE) is 8.85544