Research On Personnel Security Risk Assessment Based On AFSA-BP Neural Network

As a modern scientific management method,safety risk assessment has the characteristics of forward-looking,predictive,and preventive.It can effectively reduce the loss caused by accidental factors,scientifically and effectively reduce the probability of accidents,and ensure safety and stability.The personnel security risk assessment system plays an important role in the security and stability of the army.It is necessary to adopt scientific and effective means to conduct security assessments to prevent problems before they occur.This article first studies the relevant theories of neural networks according to the requirements of personnel security risk assessment,and then conducts a demand analysis of the system,including the main roles of using the system and the functional requirements of each role,and analyzes the system constraints,hardware environment,and software Non-functional requirements such as the environment.Then combined with the artificial fish school algorithm,the system was modeled,and the main influencing factors of the personnel safety risk assessment were extracted: the personnel’s military age,military rank,award honors and other seven key factors,and then combined with the AFSA-BP algorithm for evaluation.Then,according to the improvement strategy of the AFSA-BP evaluation algorithm,the evaluation algorithm is improved,and the evaluation model based on the virus evolution genetic algorithm is established,and the iterative convergence of the virus evolution artificial fish school algorithm is compared and analyzed through simulation experiments.Finally,the system uses Spring,Struts and Hibernate open source frameworks,Mysql database,My Eclipse development tools,and Tomcat server to build a system development environment to realize the main functions of the system,and to verify the improvement of the algorithm by experiments.It is found through data comparison and analysis.After the improvement,the efficiency of the algorithm is significantly improved,and the error between the evaluation of the personnel safety risk value and the reference value is also significantly reduced.