An efficient PAM spatial clustering algorithm based on MapReduce

Clustering analysis has been a hot area of spatial data mining for several years. With the rapid development of the spatial information technology, the amount of spatial data is growing exponentially and it makes spatial clustering of massive spatial data a challenging task. Aiming to improve the ef...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2014 22nd International Conference on Geoinformatics s. 1 - 6
Hlavní autoři: Jun Yue, Shanjun Mao, Mei Li, Xuesen Zou
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.06.2014
Témata:
Hadoop
HBase
Joints
K-Medoids
MapReduce
PAM
Spatial Clustering
ISSN:2161-024X
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Clustering analysis has been a hot area of spatial data mining for several years. With the rapid development of the spatial information technology, the amount of spatial data is growing exponentially and it makes spatial clustering of massive spatial data a challenging task. Aiming to improve the efficiency of the clustering process on massive spatial data, an implementation of parallel Partitioning Around Medoids (PAM) spatial clustering algorithm based on MapReduce is proposed. The experiments on Hadoop and HBase demonstrate that the proposed algorithm can process massive spatial data efficiently and scale well on commodity hardware.
ISSN:2161-024X
DOI:10.1109/GEOINFORMATICS.2014.6950803