Feature selection and machine learning model optimization for DDoS detection

In this chapter, the Implementation design is divided into two sections, the system setup design, and the machine learning flow process design. Both are explained in detail in the section.
The dataset used for the experiment was obtained from the open-source database CSE-CIC-IDS2018 (‘IDS 2018’ 2022) and the generated data set.
The open-source dataset was used to train a large data set that contained 300967 instances of benign and DDoS datasets while the generated data set contained 28, 972 instances of benign and DDoS datasets. The dataset contained many fields in which “32 out of 80” features were used for the open data set and “31 out of 84” was used for the newly generated dataset.
A comparative study of supervised machine learning algorithms which will be used to predict the accuracy and clarity of how DDoS attacks are detected in the cloud will be presented and evaluated in terms of performance and accuracy.