Research on Network Security Situation Prediction and Visualization based on Security Situation Awareness

Research on network security situation prediction and visualization, grounded in security situation awareness, is paramount in today's digital landscape. This study delves into various methodologies, challenges, and future prospects in cyber situational awareness visualization. While Median Filtering is dependable, its broad approach may overlook certain abnormal values. Moving Window Average Filtering, conversely, proves adept at outlier detection, making it suitable for analyzing abnormal events accurately. Additionally, Exponential Weighted Moving Average Filtering excels in capturing sustained abnormal behaviors and adapting to intricate data backgrounds. Finally, Kalman Filtering emerges as robust in handling prolonged abnormal events with comprehensive considerations, albeit with a slightly reduced ability to identify continuous abnormal events. An evaluation of accuracy, employing Mean Absolute Error (MAE), underscores Kalman Filtering's superiority. It exhibits a consistent decreasing trend in MAE as the sample size increases, establishing it as a potent and reliable choice for accurately predicting dynamic systems in cybersecurity applications. The study underscores the predictive nature of the models employed. The selection of a filtering method hinges on the specific requirements and characteristics of the data stream. By adeptly leveraging the strengths of various filtering techniques, organizations can bolster their security posture and effectively forecast potential threats in dynamic cyber environments.