A.L. Ramdani, H.B. Firmansyah
Pillar K-means cluster is one of clustering algorithm developed from K-means algorithm which is more effective than another similar algorithm. The advanced development of electronic data has brought two main impacts in data clustering algorithm, including how to store big data and how to process this type of data. This paper discussed the analysis and implementation of Pillar K-means algorithm in a distributed system using MapReduce Framework. We cited the existing Pillar K-Means Algorithm and implemented it by using MapReduce Framework. Various functions had been implemented simultaneously such as Mapper and Reducer, which are part of MapReduce Framework. The result showed that there was a positive performance regarding efficiency and scalability of Pillar K-means by using synthetic datasets. © Published under licence by IOP Publishing Ltd.
Department of Informatic, Institut Teknologi Sumatera, Indonesia