Research on Data Link Knowledge Graph Technology and Its Application Based on Data

Lidong Zhang, Dan WangAtlantis 2024 · DOI · OpenAlex

Abstract

This paper first analyzes the dilemma of digital construction of data link equipment under the current digital background of our army and the necessity of studying the knowledge graph of data link. Then it introduces the current research status and development process of knowledge graph technology at home and abroad, combs the definition, construction, storage and application of knowledge graph, and the update and maintenance of the whole life cycle, and expounds the relationship between knowledge graph and ontology. On this basis, combined with the characteristics of data link equipment, this paper puts forward the construction method of data link knowledge graph based on data, which is of great significance for the later digital construction of our army's data link equipment, improving the effect of data governance, improving the efficiency of equipment data use, and assisting the decision-making of relevant data link practitioners.

Introduction

Modern warfare has transitioned into an information-centric model where data link equipment—the backbone of inter-platform communication—plays a critical role. However, the digital construction of this equipment faces a "data dilemma." Data is often siloed across asynchronously developed platforms, high-dimensional and difficult for commanders to interpret under pressure, and frequently of low quality due to unstructured storage. This paper, published in the proceedings of ICAID 2024, proposes a Knowledge Graph (KG) framework specifically tailored for military data link equipment to transform these fragmented data points into a structured, visual, and actionable knowledge base.

Methodology: Data Classification and Extraction

The construction process begins with a dual-axis classification of data: by "sovereign attribute" (Internal vs. External) and "storage property" (Structured vs. Unstructured). For structured data, the framework utilizes RDF triples. For unstructured data, such as technical manuals and reports, the authors employ an enhanced NLP pipeline.

A key innovation is the modification of the TF-IDF algorithm. While standard models often filter out low-frequency words, this framework uses a domain-specific "keyword dictionary" to ensure critical but rare terms like "waveform" or "time slot" are prioritized. Furthermore, the "Mode Layer" or schema serves as the logical foundation of the graph. By linking project serial numbers to capability requirements and index construction, the system creates a traceable path from high-level strategic goals to specific hardware components.

Data Link Knowledge Graph Mode Layer Schema

System Architecture: The Three-Tier Process

The proposed technical architecture is divided into three functional layers that guide the data from raw input to a refined knowledge graph:

  1. Knowledge Extraction Layer: Converts raw text and tables into semantic triples.
  2. Knowledge Fusion Layer: Employs Bayesian networks and semantic similarity algorithms to align entities and de-duplicate information from disparate sources.
  3. Knowledge Processing Layer: Focuses on quality assessment, reasoning, and visualization via a Flask-based web interface using Echarts.
Knowledge Graph Construction Process

Key Insights and Results

The research highlights that military KGs require a shift from frequency-based importance to domain-relevance. A single mention of a specific equipment defect can be more critical than a frequent generic term. While the paper focuses on architectural design, it references baseline accuracies of approximately 87% for entity recognition in similar military contexts. The integration of expert prior knowledge ensures that the "Mode Layer" remains constrained to valid military logic, improving the reliability of the resulting graph.

Conclusion

The proposed Data Link Knowledge Graph framework offers a blueprint for improving data governance and decision-making efficiency in military equipment management. Future work will focus on enhancing the real-time update capabilities of the graph and reducing the manual effort required for initial dictionary construction, ensuring the system can scale alongside evolving military technology.

Summary updated

Abstract

This paper first analyzes the dilemma of digital construction of data link equipment under the current digital background of our army and the necessity of studying the knowledge graph of data link.Then it introduces the current research status and development process of knowledge graph technology at home and abroad, combs the definition, construction, storage and application of knowledge graph, and the update and maintenance of the whole life cycle, and expounds the relationship between knowledge graph and ontology.On this basis, combined with the characteristics of data link equipment, this paper puts forward the construction method of data link knowledge graph based on data, which is of great significance for the later digital construction of our army's data link equipment, improving the effect of data governance, improving the efficiency of equipment data use, and assisting the decision-making of relevant data link practitioners.
BibTeX
@inproceedings{zhang2024research,
  author = {Lidong Zhang and Dan Wang},
  title = {{Research on Data Link Knowledge Graph Technology and Its Application Based on Data}},
  booktitle = {Atlantis},
  year = {2024},
  doi = {10.2991/978-94-6463-490-7\_57},
}
Figure