Campus Network Security Situation Awareness Based on AHP and Nadam Algorithm

Background: the campus network serves as a basic communication and management platform of colleges and universities, providing convenience for campus life. However, it also faces network security issues. Aiming at the security problems brought by network threat attacks, a security situational awareness model of campus network based on Analytic Hierarchy Process (AHP) and Nadam algorithm was proposed. Methods: firstly, the improved AHP was used to build the Network Security Situation (NSS) assessment mode. Then, the Nadam algorithm and the improved Long Short-Term Memory (LSTM) network were used to build the NSS prediction model. Results: the results showed that the improved AHP had a good consistency in the Judgment Matrix (JM). The Fuzzy Neural Network (FNN) evaluation method, based on the improved Gravity Search Algorithm (GSA), began to converge around the 69th iteration, with a small output error of 0.0107. After 20 iterations, the fitness value stabilized at 0.13. The NSS assessment model, based on the improved AHP, achieved a high security situation value of 0.425. The mean square error of the Look ahead method, combined with the Nadam algorithm, flattened out after 80 iterations, which could increase the convergence speed of LSTM networks. The accuracy of the NSS prediction model using Nadam algorithm and improved LSTM network was the highest, up to 98%. The false positive rate and false negative rate were the lowest, at 2.64% and 11.03%, respectively. Additionally, the predicted NSS value was closest to the true value, with a Mean Absolute Percentage Error (MAPE) of 0.039 and a mean square error of 0.01. Conclusion: in summary, the constructed model in this study has good application effects in NSS awareness, and has certain positive significance for maintaining the campus network security