Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure
Intelligent informationization has brought great convenience to people’s life. However, the renewal of intelligent devices makes network traffic show blowout growth, which brings more difficulties to network security anomaly detection. As a result, the study suggests a model for detecting...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10747346/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846159829766242304 |
|---|---|
| author | Guiling Fan Shaojun Chen |
| author_facet | Guiling Fan Shaojun Chen |
| author_sort | Guiling Fan |
| collection | DOAJ |
| description | Intelligent informationization has brought great convenience to people’s life. However, the renewal of intelligent devices makes network traffic show blowout growth, which brings more difficulties to network security anomaly detection. As a result, the study suggests a model for detecting anomalies in network security that is based on the stabilized layer normalized attention unit structure. The multi-layer stacked unit structure is made simpler using the intermediate layer feature distillation method. The experimental results revealed that the stable layer normalized attention unit structure could effectively prevent the occurrence of floating point overflow phenomenon in the process of data processing, and only about 620 rounds to achieve convergence, distillation structure of the student neural network fitting degree reaches more than 98%. In addition, this structure can effectively reduce the memory occupation, shorten the inference time, and the optimization rate is as high as 75%. The accuracy of the proposed model reached 98.34%, the recall was higher than 97.00%, the false alarm rate was only 2.38% on average, and the F1 value reached 97.38% on average. In summary, the proposed model in the study can not only timely and accurately detect abnormal behaviors in the network to improve network security protection capability, but also complement and enhance existing network security technology, promoting the development of the network security field. |
| format | Article |
| id | doaj-art-8e996da338704e5abd8c23bfad35795c |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-8e996da338704e5abd8c23bfad35795c2024-11-23T00:00:29ZengIEEEIEEE Access2169-35362024-01-011217200417201710.1109/ACCESS.2024.349424210747346Design of Network Security Anomaly Detection Model Based on SLNA Cell StructureGuiling Fan0Shaojun Chen1https://orcid.org/0009-0006-3369-4426College of Computer, Xijing University, Xi’an, ChinaCollege of Computer, Xijing University, Xi’an, ChinaIntelligent informationization has brought great convenience to people’s life. However, the renewal of intelligent devices makes network traffic show blowout growth, which brings more difficulties to network security anomaly detection. As a result, the study suggests a model for detecting anomalies in network security that is based on the stabilized layer normalized attention unit structure. The multi-layer stacked unit structure is made simpler using the intermediate layer feature distillation method. The experimental results revealed that the stable layer normalized attention unit structure could effectively prevent the occurrence of floating point overflow phenomenon in the process of data processing, and only about 620 rounds to achieve convergence, distillation structure of the student neural network fitting degree reaches more than 98%. In addition, this structure can effectively reduce the memory occupation, shorten the inference time, and the optimization rate is as high as 75%. The accuracy of the proposed model reached 98.34%, the recall was higher than 97.00%, the false alarm rate was only 2.38% on average, and the F1 value reached 97.38% on average. In summary, the proposed model in the study can not only timely and accurately detect abnormal behaviors in the network to improve network security protection capability, but also complement and enhance existing network security technology, promoting the development of the network security field.https://ieeexplore.ieee.org/document/10747346/Anomaly detectionattentionflow dataknowledge distillationnetwork security |
| spellingShingle | Guiling Fan Shaojun Chen Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure IEEE Access Anomaly detection attention flow data knowledge distillation network security |
| title | Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure |
| title_full | Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure |
| title_fullStr | Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure |
| title_full_unstemmed | Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure |
| title_short | Design of Network Security Anomaly Detection Model Based on SLNA Cell Structure |
| title_sort | design of network security anomaly detection model based on slna cell structure |
| topic | Anomaly detection attention flow data knowledge distillation network security |
| url | https://ieeexplore.ieee.org/document/10747346/ |
| work_keys_str_mv | AT guilingfan designofnetworksecurityanomalydetectionmodelbasedonslnacellstructure AT shaojunchen designofnetworksecurityanomalydetectionmodelbasedonslnacellstructure |