Related Experiment Video
Updated: Jun 21, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.0K
A multi-information fusion anomaly detection model based on convolutional neural networks and AutoEncoder
Zhongnan Zhao1,2, Hongwei Guo3, Yue Wang4
1School of Computer Science and Technology, Harbin University of Science and Technology, Harbin, 150080, China. zhaozhongnan@hrbust.edu.cn.
Scientific Reports
|July 12, 2024
Summary
This study introduces a novel multi-information fusion model for network traffic anomaly detection. The model enhances security by combining convolutional neural network and AutoEncoder features for more accurate identification of network threats.
Area of Science:
- Computer Science
- Cybersecurity
- Network Engineering
Background:
- Network traffic anomaly detection is crucial for maintaining secure operations in dynamic network environments.
- Existing methods can suffer from information loss during traffic data processing, impacting detection accuracy.
- Effective feature extraction is key to improving network security analysis.
Purpose of the Study:
- To propose a multi-information fusion model for enhanced network traffic anomaly detection.
- To address information loss issues in traffic data analysis.
- To improve the overall performance and accuracy of network security detection systems.
Main Methods:
- A novel model integrating a convolutional neural network (CNN) and an AutoEncoder (AE) was developed.
- CNN extracts features directly from raw network traffic data.
- AE encodes statistical features to supplement information loss, creating a fused feature representation.
Main Results:
- The proposed model effectively combines raw traffic data features with global statistical features.
- This integrated feature representation provides a more complete view of network traffic information.
- Experimental results demonstrate superior classification accuracy compared to classical machine learning methods.
Conclusions:
- The multi-information fusion model significantly improves network traffic anomaly detection performance.
- This approach offers a robust solution for identifying anomalies in complex network environments.
- The findings contribute to advancing network security through improved data analysis techniques.
