基于改进GWO算法的多源物联网设备识别研究
Research on Multi-Source IoT Device Recognition Based on Improved GWO Algorithm
张浩 1孙力1
作者信息
- 1. 兰州工商学院 信息工程学院,甘肃 兰州 730101
- 折叠
摘要
物联网设备的不断增多,导致物联网的数据容量存储和数据安全面临着极大的挑战.为了实时监测物联设备,提供更好的安全性服务.在介绍支持向量机分类器的基础上引入了灰狼优化算法进行结合,同时接入余弦值变换定律进行参数优化,然后以集成学习的方式进行了多分类器组合,并构建了一种新型设备识别模型.实验结果表明,改进灰狼优化算法的类准确率最高可达95%,迭代次数最少为1180 次,精准度最高为93.4%,召回率值最高为92.8%,F1 值最高为92.9%.在7 个设备的识别测试中,能够完成90 分以上的识别效果.由此可知,所改进的灰狼优化算法和最终的识别模型都具有较高的鲁棒性和优越性,能够为物联网设备识别领域的技术发展提供一定的技术支持.
Abstract
The increasing number of IoT devices has led to great challenges in data capacity storage and data security in IoT.In order to monitor IoT devices in real time and provide better security services.The study first introduces the gray wolf optimization algorithm for combination based on the introduction of support vector ma-chine classifiers,while accesses the cosine value transformation law for parameter optimization.Then a multi-classifier combination is carried out with integrated learning and a novel device recognition model is constructed.The experimental results show that the class accuracy of the improved gray wolf optimization algorithm is up to 95%,the minimum number of iterations is 1180,the precision is up to 93.4%,the recall value is up to 92.8%,and the F1 value is up to 92.9%.7 devices are able to accomplish more than 90 points in the recognition test.It can be seen that the improved gray wolf optimization algorithm and the final recognition model of the in-stitute have high robustness and superiority,and can provide certain technical support for the technical develop-ment in the field of IoT device recognition.
关键词
物联网/设备识别/灰狼优化算法/支持向量机/余弦值变化Key words
Internet of Things/device identification/gray wolf optimization algorithm/support vector ma-chine/cosine value variation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据