基于半监督学习的无线信道场景识别
Wireless Channel Scenario Classification Based on Semi-supervised Learning
谭思源1
作者信息
- 1. 西安电子工程研究所,陕西 西安 710000
- 折叠
摘要
为了改善有监督学习的泛化性较差,只能较好地识别已经见过的用于训练的信道数据属于哪种信道场景的问题,文章提出了一种基于伪标签半监督学习方法的无线信道场景识别方法,仿真结果表明,在识别新的信道数据(来源不同但属于模型中的某一类信道场景)所对应的信道场景时,半监督学习方法的识别准确率远高于有监督学习方法的识别准确率.由此可见,半监督学习的方法可以提高无线信道场景识别模型的泛化能力.
Abstract
To address the issue of poor generalization of supervised learning,which can only effectively classify which channel scenario the channel data used for training belongs to,this paper proposes a wireless channel scenario classification method based on pseudo-label semi-supervised learning.Simulation results indicate that,when classifying the channel scenario corresponding to new data(originating from different sources but belonging to a known category of channel scenario in the model),the semi-supervised learning approach significantly outperforms supervised learning in terms of classification accuracy.Thus it can be seen,it is concluded that semi-supervised learning can enhance the generalization ability of wireless channel scenario classification models.
关键词
信道场景识别/半监督/伪标签Key words
channel scenario classification/semi-supervised learning/pseudo label引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
维普
万方数据