A Network Security Situational Assessment Method Considering Spatio-Temporal Correlations

Symmetry, Journal Year: 2025, Volume and Issue: 17(3), P. 385 - 385

Published: March 3, 2025

Network security situational assessment is crucial for network monitoring and management. Existing methods often fail to consider spatio-temporal correlations, limiting their accuracy. This paper proposes a method that integrates these correlations improved assessment. The first addresses the challenges posed by numerous nodes large time-series data designing an anomaly detection approach based on state fluctuations symmetry. It filters time window identify key symmetrical patterns, reducing computational overhead. Next, metric developed single window, incorporating both temporal spatial components. Temporal measures between consecutive windows, while identifies four types of abnormal situations. Finally, results across windows are aggregated, considering historical current events. Historical event impacts attenuated using decay function, events weighted progression stage. Experiments multiple datasets validate method’s effectiveness reasonableness in assessing security. average execution BP 3.8987 s. proposed 0.2117 s, saving 3.687 s compared method. LSTM (Long Short-Term Memory) 0.9427 2.956 method, but it still 0.731 slower than

Language: Английский

---

A novel ensemble deep reinforcement learning model for short‐term load forecasting based on Q‐learning dynamic model selection

The Journal of Engineering, Journal Year: 2024, Volume and Issue: 2024(7)

Published: July 1, 2024

Abstract Short‐term load forecasting is critical for power system planning and operations, ensemble methods electricity loads have been shown to be effective in obtaining accurate forecasts. However, the weights prediction models are usually preset based on overall performance after training, which prevents model from adapting face of different scenarios, limiting improvement performance. In order improve accurateness validity method further, this paper proposes an deep reinforcement learning approach using Q‐learning dynamic weight assignment consider local behaviours caused by changes external environment. Firstly, variational mode decomposition used reduce non‐stationarity original data decomposing sequence. Then, recurrent neural network (RNN), long short‐term memory (LSTM), gated unit (GRU) selected as basic predictors. Finally, optimal ensembled three sub‐predictors generated algorithm, final results obtained combining their respective predictions. The show that capability proposed outperforms all sub‐models several baseline methods.

Language: Английский

---

VTion-PatchTST: Elevated PatchTST model for network security situation prediction

Computers & Electrical Engineering, Journal Year: 2024, Volume and Issue: 118, P. 109393 - 109393

Published: June 19, 2024

Language: Английский

---

Prediction of Network Security Situation Based on Attention Mechanism and Convolutional Neural Network–Gated Recurrent Unit

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(15), P. 6652 - 6652

Published: July 30, 2024

Network-security situation prediction is a crucial aspect in the field of network security. It primarily achieved through monitoring behavior and identifying potential threats to prevent respond attacks. In order enhance accuracy prediction, this paper proposes method that combines convolutional neural (CNN) gated recurrent unit (GRU), while also incorporating an attention mechanism. The model can simultaneously handle spatial temporal features optimize weight allocation Firstly, CNN’s powerful feature extraction ability utilized extract behavior. Secondly, time-series are processed GRU layer. Finally, model’s performance further, we introduce mechanisms, which dynamically adjust importance different based on current context information; enables focus more critical information for accurate predictions. experimental results show network-security method, CNN introduces mechanism, performs well terms fitting effect effectively prediction.

Language: Английский

---

Network Traffic Prediction Mechanism Based on Variational Mode Decomposition and GA-ELM

Proceedings of the 2021 5th International Conference on Electronic Information Technology and Computer Engineering, Journal Year: 2024, Volume and Issue: unknown, P. 964 - 968

Published: Oct. 18, 2024

Language: Английский

---