Scholarlink

The rapid growth in the technology advancement has increased the demand for smart device. The edge computing applications are rising in the due to its vast benefits of storage and accessibility. As the users are increasing, the vulnerable activities are also rising in the network that cause security threats for the edge computing data. The existing methods tried to detect the security threats in the network to secure the data but did not get the expected outcome due to poor representation of node networks. The graph neural network-based long short-term memory with common vulnerability scoring system (GNN-LSTM with CVSS) model was developed for the detection of security threats in the edge computing. The graph neural network (GNN) represented the node network graph to analyze the threats. The ResNet-50 extracted the important details from the node network graph. The long short-term memory (LSTM) interpretated the patterns of the extracted features for the security threat detection in the edge computing. The developed GNN-LSTM with CVSS has increased the security threat detection performance. The proposed GNN-LSTM with CVSS model has obtained greater results of 94.27% accuracy and 93.73% f1-score compared to the existing distilled-bidirectional encoder representations from transformer (DistilBERT) model.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Graph Neural Network-Based Long Short-Term Memory with Common Vulnerability Scoring System for Security Threat Detection in Edge Computing

  • 章节

  • 发表日期：

    2026年

查看全文
  • H. Manoj T. Gadiyar,
  • K. Arjun,
  • Mohan Ramachandra Naik,
  • M. Bharathraj Kumar,
  • Gurusiddayya Hiremath

摘要

The rapid growth in the technology advancement has increased the demand for smart device. The edge computing applications are rising in the due to its vast benefits of storage and accessibility. As the users are increasing, the vulnerable activities are also rising in the network that cause security threats for the edge computing data. The existing methods tried to detect the security threats in the network to secure the data but did not get the expected outcome due to poor representation of node networks. The graph neural network-based long short-term memory with common vulnerability scoring system (GNN-LSTM with CVSS) model was developed for the detection of security threats in the edge computing. The graph neural network (GNN) represented the node network graph to analyze the threats. The ResNet-50 extracted the important details from the node network graph. The long short-term memory (LSTM) interpretated the patterns of the extracted features for the security threat detection in the edge computing. The developed GNN-LSTM with CVSS has increased the security threat detection performance. The proposed GNN-LSTM with CVSS model has obtained greater results of 94.27% accuracy and 93.73% f1-score compared to the existing distilled-bidirectional encoder representations from transformer (DistilBERT) model.