One of the typical threats to internet services’ accessibility is referred to attack of Distributed denial of service (DDoS). These attacks could be detected applying the unsupervised solutions based on ML which were trained on big datasets of DDoS attack previously for recognizing bad patterns automatically exist in incoming traffic. Here, the multiple strategy is used in the presented technique. Now, three algorithms of machine learning (ML) and red deer algorithm (RDA) will be combined together to increase accuracy of DDoS attacks detection. In this study, a experimentation need to done by applying algorithms of classification for appropriately grouping DDoS attack from legitimate flows. The best performing techniques are recognized to be ensembles via classifiers of ML obtained the comparable accuracy level. In comparison to techniques, proposed method accuracy obtained 94.93% and 0.36% of increment in accuracy
