Optimization of Computer Network Security System Based on Improved Neural Network Algorithm and Data Searching
Chongfeng Tian,Zhihao Chen,Yi Zhu,Hongfei Lu,Guoxiao Li,Rongquan Li,Wei Pan
DOI: https://doi.org/10.12694/scpe.v25i5.3150
2024-01-01
Abstract:In the realm of computer network security, an escalating need for robust and adaptive systems prompts the development of innovative approaches. This paper introduces a novel framework, termed ”ALPSO AutoLSTM-PSO Security Optimization Framework,” designed for the optimization of computer network security systems. The framework synergistically integrates advanced techniques, including Autoencoder (Auto), Long Short-Term Memory (LSTM), and Particle Swarm Optimization (PSO). The Autoencoder, trained on normal network traffic data, serves as a feature learning mechanism, capturing essential representations. The LSTM, adept at modeling temporal dependencies, complements this by recognizing sequential patterns in network behavior. Furthermore, the PSO algorithm is employed to finely tune the parameters of both the Autoencoder and LSTM networks, enhancing their collective performance. The integrated model, forged through this holistic approach, forms the cornerstone of an improved neural network algorithm. To demonstrate the efficacy of the proposed ALPSO, comprehensive experiments are conducted using the NSL-KDD dataset. This dataset provides a realistic and diverse set of network traffic scenarios, enabling a thorough evaluation of the framework’s capabilities. The algorithm, enriched by the dynamic fusion of Autoencoder and LSTM outputs, is adept at anomaly detection and security threat identification. This framework, coupled with efficient data searching techniques, enables real-time analysis of network traffic, thereby fortifying the security infrastructure. The ALPSO Framework represents a comprehensive solution that amalgamates state-of-the-art technologies to address the evolving challenges in computer network security.