Presenting a model for improving security in cloud computing to prevent distributed denial of service attacks using extreme machine learning and artificial intelligence

Document Type : Original Article

Authors
amir hossein nazari afshar 1
hamidreza doost 2
1 Graduated from the Master's degree of Tehran Azad University, Yadegar Imam Khomeini Branch, Ray Shahr
2 Master's degree student in Information Technology Engineering, E-Commerce major, Azad University, Science and Research Branch.
Abstract
Several machine learning-based solutions have been proposed for detecting distributed denial of service attacks in cloud computing. This study presents a model for improving security in cloud computing to prevent distributed denial of service attacks using extreme machine learning and artificial intelligence. In this method, an improved SaE-ELM model is developed that can adapt the mutation strategy, crossover rate, and crossover operator and is able to automatically determine the appropriate number of hidden layer neurons. To evaluate the proposed method used for attack detection on the web, the OSELM algorithm was used and it was considered with several other peer methods based on different criteria for evaluation. For the proposed network, 15 neurons in the hidden layer, 2500 iterations in training and a Zigmoid kernel function were proposed and evaluated using the NSL-KDD dataset. The proposed method achieved a detection accuracy of 86.80% with NSL-XKD, and experiments showed that the performance of the proposed attack detection system is better than the original SaE-ELM-based system and advanced techniques. However, it resulted in a longer training time than the SaE-ELM-based system.
Keywords
Security
cloud computing
distributed denial of service attacks
extreme machine learning
and artificial intelligence
Subjects