提示
基于多特征融合的恶意流量检测研究
首发时间：2025-03-10
摘要：随着互联网技术的快速发展和网络规模的不断扩大，网络安全问题日益严峻，尤其是异常流量的检测。本文研究了基于机器学习的异常流量检测方法，提出了一种结合随机森林和互信息的特征融合方法（rf-mi），并应用于多种常见的机器学习模型进行比较。通过对cicids2017数据集的处理和分析，本文采用了多种特征融合方法（如随机森林、互信息、lasso）以及基于catboost的异常流量检测模型，评估了其在异常流量检测任务中的表现。实验结果表明，rf-mi方法在特征融合方面的效果显著优于其他方法，且基于catboost的模型在准确率、精确率、召回率和f1值等指标上均表现最佳。通过特征融合和模型优化，本文提出的方法不仅提高了异常流量的检测精度，还增强了网络安全防护的能力。研究结果为网络流量分析和安全检测提供了有效的凯发k8国际首页的解决方案，具有重要的理论价值和实际应用意义。）
关键词：
for information in english, please click here
research on anomalous traffic detection based on multi-feature fusion
abstract：with the rapid development of internet technology and the continuous expansion of network scale, network security issues have become increasingly severe, especially in the detection of anomalous traffic. this paper investigates machine learning-based methods for anomalous traffic detection, proposing a feature fusion method (rf-mi) that combines random forest and mutual information, and applies it to compare with various common machine learning models. through the processing and analysis of the cicids2017 dataset, this paper adopts several feature fusion methods (such as random forest, mutual information, and lasso), as well as a catboost-based anomalous traffic detection model, to evaluate their performance in the anomalous traffic detection task. the experimental results show that the rf-mi method significantly outperforms other methods in feature fusion, and the catboost-based model performs best in terms of accuracy, precision, recall, and f1-score. by integrating feature fusion and model optimization, the proposed method not only improves the detection accuracy of anomalous traffic but also enhances the network security defense capability. the research results provide an effective solution for network traffic analysis and security detection, with significant theoretical value and practical application significance.
keywords：
基金：
论文图表：
引用
导出参考文献
no.****
同行评议
共计0人参与
勘误表
基于多特征融合的恶意流量检测研究
评论
全部评论