A MapReduce based ISODATA algorithm

Author

Wan, Cong ; Wang, Cuirong ; Song, Xin

Author_Institution

Coll. of Inf. Sci. & Eng., Northeastern Univ., Shenyang, China

fYear

2012

fDate

15-17 July 2012

Firstpage

765

Lastpage

768

Abstract

Cluster analysis is a mathematical method that applied in various fields, such as biology, medicine, business and marketing. ISODATA algorithm is a Clustering algorithm that has been widely used. With the development of information technology, data set expanded dramatically which becomes a great challenge to the traditional algorithm. Parallel computing is one of the main methods to solve such problem. We have proposed a Parallel ISODATA algorithm based on MapReduce which is a famous distributed computing framework. Experiments show that it greatly improves the efficiency of the algorithm.

Keywords

data analysis; iterative methods; parallel processing; pattern clustering; MapReduce based ISODATA algorithm; cluster analysis; clustering algorithm; data set; distributed computing framework; information technology development; iterative self-organizing data analysis technique; mathematical method; parallel ISODATA algorithm; parallel computing; Algorithm design and analysis; Classification algorithms; Clustering algorithms; Educational institutions; Machine learning algorithms; Parallel processing; Standards;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Information Processing (ICICIP), 2012 Third International Conference on

Conference_Location

Dalian

Print_ISBN

978-1-4577-2144-1

Type

conf

DOI

10.1109/ICICIP.2012.6391498

Filename

6391498