6G业务场景的不完全多视图聚类分析

doi:10.19665/j.issn1001-2400.20230703

摘要/Abstract

摘要：

在6G网络中,由于业务种类繁杂且需求各不相同,5G网络中划分的三大业务场景已无法满足其粒度上的要求,这给6G按需服务目标的实现带来了巨大挑战。针对海量杂乱的6G场景和6G场景分类中业务数据量庞大以及数据缺失问题,提出了一套基于业务关键性能指标的多维度场景聚类分析方案。该方案基于不完全多视图聚类技术,在上千种参数组合下使用肘部法和轮廓系数法进行调参聚类。聚类结果表明,提出的方案能在不完整的场景数据集中保证收敛,并达到较高的轮廓系数值。此外,通过对比不同比例的缺失数据聚类实验,所提出的6G场景聚类方案能够有效完成对于不同程度缺失数据的多维度聚类。最后,结合原始数据和聚类标签,分析并提炼聚类得到了11类场景的场景知识及各场景的关键性能指标特征,从而为未来6G网络中的新兴场景及业务提供方法基础和理论参考。

关键词: 6G, 场景聚类, 关键性能指标, 不完全多视图聚类

Abstract:

In the 6G network,due to the variety of business types and different requirements,the three major business scenarios divided in the 5G network can no longer meet the granularity requirements,which brings great challenges to the realization of the goal of 6G on-demand services.Aiming at the massive and messy 6G scenarios and the huge amount of business data and data missing in the classification of 6G scenarios,this paper proposes a set of multi-dimensional scenario clustering analytical schemes based on business key performance indicators.The scheme is based on the incomplete multi-view clustering technology,and uses the elbow method and the silhouette coefficient method to perform parameter tuning clustering under thousands of parameter combinations.Clustering results show that the scheme proposed in this paper can guarantee convergence in incomplete scene datasets and achieve high silhouette coefficient values.In addition,by comparing the missing data clustering experiments with different proportions,the proposed 6G scene clustering scheme can effectively complete the multi-dimensional clustering for different degrees of missing data.Finally,this paper combines the original data and clustering labels,analyzes and refines the clustering to obtain the scene knowledge of 11 types of scenarios and the characteristics of key performance indicators of each scenario,so as to provide the method basis and theoretical reference for emerging scenarios and services in the future 6G network.

Key words: 6G, scene clustering, KPI, incomplete multi-view clustering

中图分类号:

TN92

张茹倩, 承楠, 陈文, 李长乐. 6G业务场景的不完全多视图聚类分析[J]. 西安电子科技大学学报, 2024, 51(3): 76-87.

ZHANG Ruqian, CHENG Nan, CHEN Wen, LI Changle. Incomplete multi-view clustering analysis of 6G business scenarios[J]. Journal of Xidian University, 2024, 51(3): 76-87.

图/表 11

图1

图2

表1

图3

表2

图4

表3

轮廓系数值"

$m μ$	k	2×k	3×k	4×k	5×k	6×k	7×k
0.001	0.868 4(k=13)	0.996 8(k=11)	0.996 0(k=10)	0.970 2(k=7)	0.927 0(k=5)	0.841 6(k=10)	0.824 4(k=11)
0.1	0.450 0(k=10)	0.528 4(k=15)	0.868 6(k=16)	0.833 5(k=14)	0.891 9(k=17)	0.856 2(k=17)	0.987 2(k=7)

表3

图5

图6

图7

图8

参考文献 15

[1]	NGMN Alliance. 5G White Paper[R]. San Francisco: Next Generation Mobile Networks, 2015.
[2]	崔春风, 王森, 李可, 等. 6G愿景、业务及网络关键性能指标[J]. 北京邮电大学学报, 2020, 43(6):10-17. doi: 10.13190/j.jbupt.2020-160
	CUI Chunfeng, WANG Sen, LI Ke, et al. 6G Vision,Service and Network Key Performance Indicators[J]. Journal of Beijing University of Posts and Telecommunications, 2020, 43(6):10-17. doi: 10.13190/j.jbupt.2020-160
[3]	金宁, 王庆扬. 基于聚类算法的6G典型应用场景研究[J]. 电信科学, 2022, 38(1):121-131. doi: 10.11959/j.issn.1000-0801.2022013
	JIN Ning, WANG Qingyang. Research on 6G Typical Application Scenarios Based on Clustering Algorithm[J]. Telecommunications Science, 2022, 38(1):121-131. doi: 10.11959/j.issn.1000-0801.2022013
[4]	杨艳, 李福昌, 张忠皓. 6G通感传算融合需求分析与关键技术研究[J]. 无线电通信技术, 2023, 49(1):83-88.
	YANG Yan, LI Fuchang, ZHANG Zhonghao. Requirement Analysis and Key Technology Research of 6G Synaesthesia,Communication and Computing Fusion[J]. Radio Communication Technology, 2023, 49(1):83-88.
[5]	YOU X, WANG C, HUANG J, et al. Towards 6G Wireless Communication Networks:Vision,Enabling Technologies,and New Paradigm Shifts[J]. SCIENCE CHINA Information Sciences,2020, 020300(2016):1-76.
[6]	GARCÍA-ALONSO C R, PÉREZ-NARANJO L M, FERNÁNDEZ-CABALLERO J C. Visualizing Data using t-SNE Laurens[J]. Annals of Operations Research, 2014, 219(1):187-202.
[7]	SINGH A K, SHASHI M. Vectorization of Text Documents for Identifying Unifiable News Articles[J]. International Journal of Advanced Computer Science and Applications, 2019, 10(7):305-310.
[8]	LIU S, LIU X, WANG S, et al. Fast Incomplete Multi-View Clustering with View-Independent Anchors[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022:1-12.
[9]	YANG Y, WANG H. Multi-View Clustering:A Survey[J]. Big Data Mining and Analytics, 2018, 1(2):83-107.
[10]	WANG S, LIU X, LIU L, et al. Highly-Efficient Incomplete Large-Scale Multi-View Clustering with Consensus Bipartite Graph[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2022:9776-9785.
[11]	LI X, ZHANG H, WANG R. Multiview Clustering:A Scalable and Parameter-Free Bipartite Graph Fusion Method[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(1):330-344.
[12]	SYAKUR M A, KHOTIMAH B K, ROCHMAN E M S, et al. Integration K-Means Clustering Method and Elbow Method For Identification of The Best Customer Profile Cluster[C]//IOP Conference Series: Materials Science and Engineering. Bristol: IOP Publishing, 2018, 336:012017.
[13]	ARANGANAYAGI S, THANGAVEL K. Clustering Categorical Data Using Silhouette Coefficient as a Relocating Measure[C]//Proceedings International Conference on Computational Intelligence and Multimedia Applications,ICCIMA. Piscataway:IEEE, 2007:97-102.
[14]	LI Z, TANG C, ZHENG X, et al. High-Order Correlation Preserved Incomplete Multi-View Subspace Clustering[J]. IEEE Transactions on Image Processing, 2022, 31:2067-2080. doi: 10.1109/TIP.2022.3147046 pmid: 35188891
[15]	秦宁宁, 张臣臣. 模糊聚类下的接入点选择匹配定位算法[J]. 西安电子科技大学学报, 2022, 49(4):71-81.
	QIN Ningning, ZHANG Chenchen. Access Point Slection Matching Positioning Algorithm under Fuzzy Clustering[J]. Journal of Xidian University, 2022, 49(4):71-81.

参数	含义
n	业务样本数量
V	视图数量(KPI种类)
k	聚类簇数
m	选择的锚点数量
$\boldsymbol{\alpha} \in \mathbf{R}^{V \times 1}$	视图的权重系数向量
d_p	第p个视图的维度
$\boldsymbol{X}_{p} \in \mathbf{R}^{d}{_{p} \times n}$	第p个视图的数据矩阵
$\boldsymbol{H}_{p} \in\{0,1\}^{n \times n_{p}}$	第p个视图数据的{0,1}指标矩阵
$\boldsymbol{B}_{p} \in \mathbf{R}^{d}{_{p}}^{\times{m}}$	第p个视图的锚点矩阵
$\boldsymbol{Z}_{p} \in \mathbf{R}^{m \times n}$	第p个视图的锚图
$\boldsymbol{Z} \in \mathbf{R}^{m \times n}$	所有数据视图的共识锚图

样本特征(KPI)	速率	时延	可靠性	移动性	连接密度	覆盖范围	定位精度	可用性
有数据的样本数	335	420	300	262	111	259	330	279
向量化维度	861	975	783	699	705	956	981	954
缺省数据样本数	154	69	189	227	378	230	159	210
数据缺省率/%	31.5	14.1	38.7	46.4	77.3	47	32.5	42.9