🤖 AI Summary
Existing stroke-domain knowledge graphs (KGs) suffer from insufficient accuracy and semantic depth. Method: This paper proposes a novel KG construction framework integrating large language models (LLMs) with mathematical modeling. GPT-4 is systematically embedded across all KG construction stages—literature preprocessing, entity-relation extraction, and embedding generation—complemented by expert validation to ensure quality. A mathematical optimization model refines graph topology and enforces semantic consistency. Contribution/Results: The approach achieves state-of-the-art performance with precision (0.923) and recall (0.918), outperforming Wikidata and WN18RR benchmarks. The resulting high-quality stroke-domain KG comprises 2,692 nodes (13 types) and 5,012 edges (24 types). In the fully automated construction phase, it attains precision of 0.906 and recall of 0.923.
📝 Abstract
The purpose of this study is to introduce SKG-LLM. A knowledge graph (KG) is constructed from stroke-related articles using mathematical and large language models (LLMs). SKG-LLM extracts and organizes complex relationships from the biomedical literature, using it to increase the accuracy and depth of KG in stroke research. In the proposed method, GPT-4 was used for data pre-processing, and the extraction of embeddings was also done by GPT-4 in the whole KG construction process. The performance of the proposed model was tested with two evaluation criteria: Precision and Recall. For further validation of the proposed model, GPT-4 was used. Compared with Wikidata and WN18RR, the proposed KG-LLM approach performs better, especially in precision and recall. By including GPT-4 in the preprocessing process, the SKG-LLM model achieved a precision score of 0.906 and a recall score of 0.923. Expert reviews further improved the results and increased precision to 0.923 and recall to 0.918. The knowledge graph constructed by SKG-LLM contains 2692 nodes and 5012 edges, which are 13 distinct types of nodes and 24 types of edges.