郑州大学学报（理学版）2026，Vol.58Issue(3)：41-49,9.DOI:10.13705/j.issn.1671-6841.2024176
基于GCN的IOTA寄生链检测
Using GCN to Detect Parasitic Chains Attacks in IOTA
摘要
Abstract
With the deveopment of the Internet of Things(IoT)technology,transaction security in IoT environments become critical.The IOTA network,a distributed ledger technology designed specifically for IoT,plays a crucial role in managing transactions among numerous devices.Parasitic chain attacks,a common threat,undermine network security and performance by validating illegal transactions in the IOTA main tangle.A method based on graph convolutional networks(GCN)was proposed to detect parasitic chain transactions in the IOTA network.By analyzing the behavioral differences between normal and par-asitic chain transactions,four attributes were identified as feature values to capture these differences.Simulated IOTA networks containing parasitic chains were constructed based on the attack rules to gener-ate datasets,and the trained model was used to classify and identify malicious nodes.Experimental re-sults demonstrated that the model achieved over 80%accuracy in detecting malicious transactions,effec-tively identifying parasitic chain transactions in the network.关键词
区块链/IOTA/寄生链攻击/图卷积神经网络/物联网Key words
block chain/Internet of Things application(IOTA)/parasitic chain attack/graph convolu-tional network/Internet of Things(IOT)分类
信息技术与安全科学引用本文复制引用
刘韦淇,侯永超,木又青,丁智颖,刘明灏,赵金东..基于GCN的IOTA寄生链检测[J].郑州大学学报（理学版）,2026,58(3):41-49,9.基金项目
国家自然科学基金项目(62405262) （62405262）
