Sports Injury Knowledge Graph Construction and Intelligent Diagnosis System Based on Deep Learning

The current sports injury diagnosis process generally relies on manual experience, information is scattered, and there is a lack of systematic intelligent support, which leads to low accuracy and efficiency of diagnosis. This paper introduces a sports injury knowledge graph construction and intelligent diagnosis system method based on deep learning. First, named entity recognition and relationship extraction models are used to extract structured knowledge from multisource sports medicine data to construct a knowledge graph covering entities and relationships such as injury types, symptoms, and treatment plans. Then, graph neural networks are used to represent, learn, and reason about the graph. Finally, an intelligent diagnosis system is designed to semantically match and reason with user input information and the graph, and output diagnostic suggestions and treatment paths. Experimental results show that the entity recognition accuracy of the constructed knowledge graph reaches 91.3 %, the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{F 1}$</tex> value of relationship extraction is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 8. 7 \%}$</tex>, the graph reasoning accuracy reaches 85.6 %, and the output accuracy of the diagnosis system on the comprehensive test set reaches <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 9. 2 \%}$</tex>. This method effectively supports the intelligence and standardization of the diagnosis process.