Analyzing threat flow over network using ensemble-based dense network model

Cyberattacks may occur in any device with an Internet connection. The majority of businesses either advise preventative measures or creating gadgets with integrated cyber threat protection mechanisms. However, the availability of tools and methods needs to go beyond standard preventative measures which make the process more difficult to identify cyber threats. One important tool for combating these intrusions is an intrusion detection system based on deep learning. To analyze intrusion detection systems, this study suggests random forest-based ensemble methods. Using random forest, tests were carried out in the first phase. In the subsequent stage, random forest is utilized due to their recent notable advancements in supervised learning performance. Deep learning methods like long short-term memory (LSTM) and autoencoder (AE) networks are used in the experiment. The work is optimized using Harris hawks optimization (HHO). For experimental purposes, the Kaggle dataset is utilized. Using this dataset, the results demonstrate that IDS have greatly improved, surpassing the state of the art. The applicability model in IDS is strengthened by this enhancement.