Comparison of Different Partition Clustering Algorithms under the Network Flow Detection Scenario
哲 阿部, Shuaichen Ye · 2024
The clustering algorithm has always been considered as an effective manner to address the network flow filtering and detecting issue. Among all clustering algorithms, partition clustering is the most extensively used due to its numerous advantages such as simplicity and scalability. Many researches have made efforts on improving the performance of the partition clustering algorithm to adapt different application scenarios, thus several categories are generated. However, research regarding systematic comparison among different categories under the same scenario is still lacking, which brings uncertainty to the network flow detection. Therefore, this paper selects four typical partition clustering algorithms as examples and compares their performance under the same dataset. Some conclusions can be drawn as reference for network security professionals to carry out preliminary traffic analysis.