Scholarlink

<p>Network analysis has been widely utilized across various research domains, however most studies have focused solely on singular node types and their associations within networks, overlooking complex relationships between heterogeneous nodes. These different types of nodes and their relationships contain rich information. Network security research involves network physical security, information security and related laws and regulations. Therefore, the field of network security is a typical interdisciplinary and cross-professional discipline. Existing research analyses in the field of network security often only provide a one-sided perspective. Thus, constructing heterogeneous networks encompassing diverse nodes and complex interrelations, and leveraging these networks to extract meaningful insights, presents a key challenge. This study proposes a method to construct a heterogeneous multi-layer network P4OE model. Firstly, the homogeneous network of fully connected cooperative network is constructed for each type of personnel nodes and content nodes with cooperative relationship, and the homogeneous network of bipartite network is constructed between different relationships among papers, personnel, institutions, locations and contents. Then, through the correlation between different types of nodes, the homogeneous networks are fused into heterogeneous multi-layer networks, and the network information entropy is added to evaluate the network information structure. This paper collects the fund achievements of the open source network security theme from 1995 to 2023, and constructs a heterogeneous network containing five types of nodes and their relationships. The Leiden community detection algorithm is utilized to identify the major research areas in the network. The associations and subject evolution trends between the top ten different areas are analyzed. The centrality and PageRank values of nodes across different areas are calculated to analyze the most influential nodes and types. Finally, disciplinary cross-integration between domains is examined via cosine similarity and principal component analysis. This approach enables constructing heterogeneous networks with diverse nodes and intricate interconnections, uncovering research hotspots and trends, pinpointing key researchers and institutions, and probing cross-domain integration through network analytics, thereby enhancing research efficiency.</p>

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

Exploring the evolution trends of network security research field based on P4OE heterogeneous network

  • 期刊论文

  • 发表日期：

    2026年3月14日

查看全文
  • Yanfei Liu,
  • Dezhi Liu,
  • Wenjun Wang,
  • Chengbo Yu

摘要

Network analysis has been widely utilized across various research domains, however most studies have focused solely on singular node types and their associations within networks, overlooking complex relationships between heterogeneous nodes. These different types of nodes and their relationships contain rich information. Network security research involves network physical security, information security and related laws and regulations. Therefore, the field of network security is a typical interdisciplinary and cross-professional discipline. Existing research analyses in the field of network security often only provide a one-sided perspective. Thus, constructing heterogeneous networks encompassing diverse nodes and complex interrelations, and leveraging these networks to extract meaningful insights, presents a key challenge. This study proposes a method to construct a heterogeneous multi-layer network P4OE model. Firstly, the homogeneous network of fully connected cooperative network is constructed for each type of personnel nodes and content nodes with cooperative relationship, and the homogeneous network of bipartite network is constructed between different relationships among papers, personnel, institutions, locations and contents. Then, through the correlation between different types of nodes, the homogeneous networks are fused into heterogeneous multi-layer networks, and the network information entropy is added to evaluate the network information structure. This paper collects the fund achievements of the open source network security theme from 1995 to 2023, and constructs a heterogeneous network containing five types of nodes and their relationships. The Leiden community detection algorithm is utilized to identify the major research areas in the network. The associations and subject evolution trends between the top ten different areas are analyzed. The centrality and PageRank values of nodes across different areas are calculated to analyze the most influential nodes and types. Finally, disciplinary cross-integration between domains is examined via cosine similarity and principal component analysis. This approach enables constructing heterogeneous networks with diverse nodes and intricate interconnections, uncovering research hotspots and trends, pinpointing key researchers and institutions, and probing cross-domain integration through network analytics, thereby enhancing research efficiency.