Blockchain network layer anomaly traffic detection method based on multiple classifier integration
To improve the comprehensive generalized feature perception ability of mixed attack traffic on the blockchain network layer, and enhance the performance of abnormal traffic detection, a blockchain layer traffic anomaly detection method was proposed that supported the comprehensive judgement of data...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Department of Journal on Communications
2023-03-01
|
| Series: | Tongxin xuebao |
| Subjects: | |
| Online Access: | http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2023066/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841540085359575040 |
|---|---|
| author | Qianyi DAI Bin ZHANG Song GUO Kaiyong XU |
| author_facet | Qianyi DAI Bin ZHANG Song GUO Kaiyong XU |
| author_sort | Qianyi DAI |
| collection | DOAJ |
| description | To improve the comprehensive generalized feature perception ability of mixed attack traffic on the blockchain network layer, and enhance the performance of abnormal traffic detection, a blockchain layer traffic anomaly detection method was proposed that supported the comprehensive judgement of data anomaly with a strong generalisation capability.Firstly, to expand the difference of the input feature subset of the base classifier used, a feature subset selection algorithm based on discrimination degree and redundant information was proposed, and the output of high sensitivity subset terms was stimulated during the feature screening process, while the generation of redundant information was suppressed.Then, the stochastic variance reduction gradient algorithm was introduced into the bagging integration algorithm to realize the dynamic adjustment of the voting weights of each base modeland improve thecapability in detecting the generalised hybrid abnormal attack traffic.Finally, LBoF algorithm was proposed to map the low-dimensional numerical vector output by the integrated algorithm to a high-dimensional space.The discrepancy of data point spatial density distribution of various samples were amplified based on the potential difference between data points to increase the recall rate of anomalous data point detection.The experimental results show that in detecting multiple hybrid attack traffic on blockchain layers, the proposed method presents an increase in the anomaly detection accuracy and recall rate, which is 1.57% and 2.71%, respectively, compared with methods based on a single classifier integration. |
| format | Article |
| id | doaj-art-f76905fc18dd44b69997ea699b76bc70 |
| institution | Kabale University |
| issn | 1000-436X |
| language | zho |
| publishDate | 2023-03-01 |
| publisher | Editorial Department of Journal on Communications |
| record_format | Article |
| series | Tongxin xuebao |
| spelling | doaj-art-f76905fc18dd44b69997ea699b76bc702025-01-14T06:23:18ZzhoEditorial Department of Journal on CommunicationsTongxin xuebao1000-436X2023-03-0144668059387621Blockchain network layer anomaly traffic detection method based on multiple classifier integrationQianyi DAIBin ZHANGSong GUOKaiyong XUTo improve the comprehensive generalized feature perception ability of mixed attack traffic on the blockchain network layer, and enhance the performance of abnormal traffic detection, a blockchain layer traffic anomaly detection method was proposed that supported the comprehensive judgement of data anomaly with a strong generalisation capability.Firstly, to expand the difference of the input feature subset of the base classifier used, a feature subset selection algorithm based on discrimination degree and redundant information was proposed, and the output of high sensitivity subset terms was stimulated during the feature screening process, while the generation of redundant information was suppressed.Then, the stochastic variance reduction gradient algorithm was introduced into the bagging integration algorithm to realize the dynamic adjustment of the voting weights of each base modeland improve thecapability in detecting the generalised hybrid abnormal attack traffic.Finally, LBoF algorithm was proposed to map the low-dimensional numerical vector output by the integrated algorithm to a high-dimensional space.The discrepancy of data point spatial density distribution of various samples were amplified based on the potential difference between data points to increase the recall rate of anomalous data point detection.The experimental results show that in detecting multiple hybrid attack traffic on blockchain layers, the proposed method presents an increase in the anomaly detection accuracy and recall rate, which is 1.57% and 2.71%, respectively, compared with methods based on a single classifier integration.http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2023066/blockchain network layerensemble learningmachine learninganomaly traffic detection |
| spellingShingle | Qianyi DAI Bin ZHANG Song GUO Kaiyong XU Blockchain network layer anomaly traffic detection method based on multiple classifier integration Tongxin xuebao blockchain network layer ensemble learning machine learning anomaly traffic detection |
| title | Blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| title_full | Blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| title_fullStr | Blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| title_full_unstemmed | Blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| title_short | Blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| title_sort | blockchain network layer anomaly traffic detection method based on multiple classifier integration |
| topic | blockchain network layer ensemble learning machine learning anomaly traffic detection |
| url | http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2023066/ |
| work_keys_str_mv | AT qianyidai blockchainnetworklayeranomalytrafficdetectionmethodbasedonmultipleclassifierintegration AT binzhang blockchainnetworklayeranomalytrafficdetectionmethodbasedonmultipleclassifierintegration AT songguo blockchainnetworklayeranomalytrafficdetectionmethodbasedonmultipleclassifierintegration AT kaiyongxu blockchainnetworklayeranomalytrafficdetectionmethodbasedonmultipleclassifierintegration |