A comprehence study of DDoS attack detecting algorithm using GRU-BWFA classifier

Abstract

Listen

Translate article

The suggested classifier, Gated Recurrent Unit Neural Network, is utilised to produce the appropriate classification in this research. It is based on Bidirectional weighted feature averaging (GRU-BWFA). The findings show that each classifier performs differently, with the accuracy rate ranging between the three. GRU-BWFA, the suggested system, has the greatest accuracy rate of 99.9%. The suggested approach outperformed current algorithms to categorise and identify several attacks from the SNMP-MIB dataset, including TCP-SYN, UDP flood, ICMP-echo, HTTP flood, Slow Loris, Slow Post, and Brute force, with a detection rate of 99.9%. As a result, the System is effectively restored in the shortest possible time.