Application and Optimization of Convolutional Neural Networks Based on Deep Learning in Network Traffic Classification and Anomaly Detection

With the rapid development of Internet technology, the complexity and diversity of network traffic have increased significantly, and traditional network traffic classification and anomaly detection methods are unable to deal with current network threats. To solve this problem, this paper proposes a network traffic classification and anomaly detection technology based on deep learning. Through the analysis and experiment of a large number of network traffic data, this paper constructs a convolutional neural network model to accurately identify and classify normal traffic and abnormal traffic. The experimental results show that the accuracy of the proposed model on the test dataset reaches 98.7%, excellent performance was achieved on the CIC-IDS2017 and ISCX VPN NOVPN datasets, with accuracies of 98.5% and 99.2%, respectively, significantly improving recall and F1 score, and effectively reducing error rates, outperforming traditional methods. In addition, this paper further optimizes the model by comparing and analyzing the performance of different network structures, and finally reduces the false alarm rate to 1.5%. This research provides effective technical support for improving network security, deeply analyzes the influence of different network structures and parameters on the performance of the model, and finally optimizes the best model, which shows strong robustness and adaptability in multiple real network environments