Title: Research on spectral clustering algorithm for network communication big data based on wavelet analysis

Authors: Xinjian Dai; Zhichao Zeng

Addresses: Basic Data Application Teaching and Research Office, Changsha Social Work College, Changsha, 410004, China ' College of Computer Science and Electronic Engineering, Hunan University, Changsha, 410082, China

Abstract: In order to classify the features of big data in network communication, improve clustering efficiency and reduce error classification rate, a spectrum clustering algorithm based on wavelet analysis is proposed. The multi-scale, one-dimensional wavelet analysis method is used to sample the network communication big data, extract the spectral feature quantity of the network communication big data, and construct the channel model of big data transmission of network communication. Combined with the fuzzy C-means clustering method, the spectral clustering is performed on network communication big data to mine association rules of the large data spectrum of network communication. Combined with wavelet decomposition method, the time-frequency conversion and feature separation of network communication big data spectrum are carried out to complete the spectrum clustering of network communication big data. Simulation results show that this method is more accurate for spectrum clustering of communication big data and improves clustering efficiency.

Keywords: wavelet analysis; network communication; big data; spectral clustering; feature extraction; clustering efficiency; fuzzy C-means clustering method; wavelet decomposition method.

DOI: 10.1504/IJAACS.2022.123459

International Journal of Autonomous and Adaptive Communications Systems, 2022 Vol.15 No.2, pp.93 - 105

Received: 20 Jun 2019
Accepted: 23 Apr 2020

Published online: 21 Jun 2022 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article