【摘要】 针对当前互联网络链路流量识别方法未考虑数据块类别,导致互联网络链路流量识别准确性和效率较低,流量识别误报率和漏报率较高的问题,提出基于负载随机性的互联网络链路流量识别方法。利用数据包负载随机性检测方法,判断网络数据流中数据块负载随机性类别。将判断结果作为机器学习方法的已标注数据集,提取数据流特征属性构成样本集合,采用ID3算法,训练分析该样本集合,寻找数据间规律,构建分类模型,通过分类模型预测和评估未知数据流量,实现互联网络链路流量识别。实验结果表明,上述方法识别不同类型流量的漏报率和误报率较低,具有较强的识别可靠性,识别时间较短,能够有效提高识别准确性和效率。更多还原

【Abstract】 Aiming at the problem that the current Internet link traffic identification method does not consider the data block category, resulting in low accuracy and efficiency of Internet link traffic identification and high false positive rate and false negative rate of traffic identification, an Internet link traffic identification method based on load randomness is proposed. The randomness detection method of packet load was used to judge the randomness category of data block load in network data flow, and the result was regarded as a labeled data set. The characteristic attributes of data flow were extracted to form a sample set. ID3 algorithm was introduced to thoroughly analyze the sample set. The rules between data were found, and the classification model was also founded. According to the classification model, the unknown data traffic was evaluated to realize the identification of Internet link traffic. The experimental results show that this method has low false negative rate and false positive rate, strong recognition reliability, short recognition time, improving accuracy and efficiency.更多还原