Dataset for An Automatic Unsupervised Complex Event Processing Rules Generation Architecture for Real-Time IoT Attacks Detection

Dataset used to perform the experiments of the article "An Automatic Unsupervised Complex Event Processing Rules Generation Architecture for Real-Time IoT Attacks Detection".

The distribution between training and testing is as follows:

Normal: 50% training, 50% testing
Subscription fuzzing: 80% training, 20% testing
Disconnection wave: 15%training, 85% testing
TCP SYN Scan: 10% training, 90% testing
UDP port Scan: 10% training, 90% testing
XMAS Scan: 10% training, 90% testing
Telnet Connection: 10% training, 90% testing

本数据集用于支撑论文《一种面向实时物联网（Internet of Things，简称IoT）攻击检测的无监督复杂事件处理（Complex Event Processing）规则自动生成架构（An Automatic Unsupervised Complex Event Processing Rules Generation Architecture for Real-Time IoT Attacks Detection）》中的相关实验。
该数据集各类样本的训练集与测试集划分比例如下：
- 正常样本（Normal）：50%用于训练，50%用于测试
- 订阅模糊攻击（Subscription fuzzing）：80%用于训练，20%用于测试
- 断开波攻击（Disconnection wave）：15%用于训练，85%用于测试
- TCP SYN扫描（TCP SYN Scan）：10%用于训练，90%用于测试
- UDP端口扫描（UDP port Scan）：10%用于训练，90%用于测试
- XMAS扫描（XMAS Scan）：10%用于训练，90%用于测试
- Telnet连接（Telnet Connection）：10%用于训练，90%用于测试