Research on Scenario Matching Method of Combat Simulation Based on BERT
WANG Zimu
GONG Jianxing
ZHANG Zhongjie
LIU Quan
Abstract：Aiming at the problem of matching between training requirements and training scenarios,a scenario matching algo-rithm called SM-BERT(Scenario Matching Based on BERT)based on BERT is proposed.SM-BERT first uses prior knowledge and unsupervised learning to extract keywords from combat scenario text.Then the field adaptation learning based on focal Loss Function is carried out for the BERT model using the battle scenario field data set.Finally,the target matching scenario is obtained according to the evaluation function of the selected scenario.The experimental results show that the accuracy of the model in scenario matching is about 87%,which shows that the model can effectively help trainees to select suitable simulation scenarios.
Keywords：combat simulation scenarioscenario matchingBERT modelkeywords extractionsemantic matching
Publication Date：2025-04-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 1031-1037 )