Design and Development of a Question-Answering System Based on Knowledge Graph for High-Altitude Aircraft Selection

Gao Yi; Hou Zihao; Zheng Miaomiao

doi:10.3969/j.issn.1004-3403.2025.10.011

Abstract：To address the issue that aircraft selection in plateau environment relies on traditional quantitative analy-sis and expert evaluation while lacking intelligent decision-making support,this paper proposes a question-answer-ing system based on knowledge graph for high-altitude aircraft selection by integrating large models and knowledge graph technologies.Firstly,multi-source heterogeneous data on high-altitude aircraft are collected and preprocessed,covering environmental parameters,aircraft performance,meteorological conditions and regulatory policies.Second-ly,through entity extraction,relationship extraction and knowledge fusion techniques,structured data are loaded in-to the Neo4j graph database to build a knowledge graph for high-altitude aircraft.Thirdly,based on the Django framework and Neo4j graph database,the system is developed with a user-friendly interface,realizing core func-tions such as knowledge querying,intelligent question-answering and graph display.Finally,the system perfor-mance is optimized through functional testing and user feedback.The innovation of this paper lies in proposing a du-al-drive architecture of"knowledge graph+lightweight large language model",which adopts the BERT-BiLSTM-CRF model for entity extraction(`F1-score of 91.2%),the ERNIE 3.0 Tiny model for intelligent question-answering generation,and combines the TransE model for knowledge completion,ultimately constructing a structured knowl-edge graph containing 897 entities and 12,000 relationships.Verified by 500 test cases,the system has achieved a question-answering accuracy of 88.6%and an F1-score of 86.3%,providing an intelligent tool for the decision-mak-ing and operational scheduling for high-altitude aircraft selection,filling the research gap in intelligent aircraft selec-tion in plateau environment.