Adaptive Asynchronous Clustering Algorithms for Wireless Mesh Networks

It is a challenge to generate an accurate machine learning model in a distributed network due to the increased concern in data privacy and high cost in gathering all raw data. This paper presents an adaptive asynchronous distributed clustering algorithm and two centralised methods for agents in wire...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on knowledge and data engineering Vol. 35; no. 3; pp. 2610 - 2627
Main Authors: Qiao, Cheng, Brown, Kenneth N., Zhang, Fan, Tian, Zhihong
Format: Journal Article
Language:English
Published: New York IEEE 01.03.2023
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Accuracy
Adaptive algorithms
Algorithms
asynchronous
Centroids
Clustering
clustering algorithm
Clustering algorithms
Computational modeling
Computer networks
Context modeling
Costs
Data models
Distributed algorithm
Finite element method
Machine learning
Network topologies
Privacy
Temperature sensors
wireless mesh network
Wireless networks
ISSN:1041-4347, 1558-2191
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is a challenge to generate an accurate machine learning model in a distributed network due to the increased concern in data privacy and high cost in gathering all raw data. This paper presents an adaptive asynchronous distributed clustering algorithm and two centralised methods for agents in wireless network to learn the global models, while the privacy is protected. Moreover, the communication cost and clustering quality can be adaptively balanced. The proposed clustering algorithms do not require the number of clusters to be pre-defined, and we propose a bounding boxes based method to fully utilize the shape information of clusters to improve the accuracy of the global model. Furthermore, we consider different knowledge levels of agents and different requirements about the global model. In experiments on randomly generated network topologies, we demonstrate that methods which do all the iterations of clustering in each cycle, and which exchange descriptions of cluster shape and density instead of just centroids and data counts, achieve higher accuracy, in significantly shorter elapsed time.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1041-4347
1558-2191
DOI:10.1109/TKDE.2021.3119550