Release Time:2019-03-09  Hits:

Indexed by: Journal Article

Date of Publication: 2015-01-01

Journal: INTERNATIONAL JOURNAL OF SENSOR NETWORKS

Included Journals: Scopus、SCIE

Volume: 18

Issue: 1-2,SI

Page Number: 32-39

ISSN: 1748-1279

Key Words: sensor networks; big sensor data; fuzzy c-means clustering; cloud computing; affinity propagation

Abstract: Fuzzy c-means (FCM) clustering has emerged as one important technique for pattern recognition, image processing and data analysis. Owing to the huge amount of data and computational complexity, FCM is difficult to cluster big sensor data in real time. The paper proposes three distributed FCM algorithms, namely distributed FCM algorithm based on MapReduce (DFCM), distributed online FCM algorithm based on sampled affinity propagation (AP) clustering (DOFCM) and distributed kernel FCM algorithm based on MapReduce (DKFCM). The proposed algorithms use cloud computing technology to improve the efficiency for clustering big sensor data. Experiments prove the superiority of the proposed methods through the clustering accuracy and clustering speed.