Design of K-means clustering algorithm in PGAS based Mapreduce framework

Adai Shomanov, Мадина Мансурова, Aliya Nugumanova

2018

Closed access Metadata (CSL-JSON)
DOI 10.1109/icaict.2018.8747118
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

Clustering large volumes of data is a complicated and time-consuming task. The main goal of clustering task is to explore a given dataset and find appropriate set of cluster centers that have maximum within-cluster and minimum inter-cluster similarity.In our work we present parallel K-means clustering algorithm based on Partitioned Global Address Space implementation of Mapreduce model. The main idea of our approach is to exploit data locality of partitioned global address space model to assign threads to different cluster centers.