Feature optimization based cyber-attack detection in public cloud network using machine learning swarm intelligence

  • Unique Paper ID: 171404
  • Volume: 11
  • Issue: 7
  • PageNo: 3420-3426
  • Abstract:
  • In current decade growth of cloud computing-based services all area of IT-enabled service sector. The growing impact of services interconnect different application and network. the interconnection of network is easy to target of intruders and cyber threats. In order to identify complex and undetected threats, the development of learning intrusion detection systems has been the main focus. Machine learning-based models are frequently used in intrusion detection systems because of their rapid accuracy gains. This study used machine learning to detect attacks on network traffic through multiple classifications. The CICIDS2017 data set, which includes both recent and older attacks, was used to build the model. Tests were run quickly on the CICIDS2017 data set, which has about 2.8 million rows of data, using the high-performance computer. Our machine learning models performed better after we cleaned, normalised, oversampled for an uneven number of labels, and used feature selection techniques to shrink the size of the data set. Using the pre-processed data set, the random forest classifier was found to have the highest accuracy of 99.94% when compared to the decision tree, logistic regression, and other classifiers.

Related Articles