Abstract
The risk assessment and prevention in traditional airport safety assurance usually rely on human experience for analysis, and there are problems such as heavy manual workload, excessive subjectivity, and significant limitations. This article proposed a risk assessment and prevention mechanism for airport security assurance that integrated LSTM algorithm. It analyzed the causes of malfunctioning flights by collecting airport flight safety log datasets. This article extracted features related to risk assessment, such as weather factors, airport facility inspections, and security check results, and conducted qualitative and quantitative analysis on these features to generate a datable risk warning weight table. This article used these data to establish an LSTM model, which trained LSTM to identify potential risks and provide early warning by learning patterns and trends in historical data. It then handed over the new data to the trained LSTM model for risk assessment and prediction, grading and warning of risks. It monitored the airport security situation in real-time based on the results and quickly notified airport security personnel to handle it. The outcome indicates that the standard error of the LSTM algorithm model training is less than 0.18, and the decision coefficients were all greater than 0.9. The predicted data was highly consistent with the actual data. It can be summarized that the algorithmic model has good accuracy and robustness. The LSTM algorithm can play a role in providing early warning, assisting decision-making, optimizing resources, and enhancing real-time monitoring in airport security assurance. It can effectively improve the safety and prevention capabilities of airports, and reduce the losses caused by potential risks.