DDoS attacks flood websites and online services with more traffic than servers and networks can handle. As technology has advanced and the Internet has become more widely used, such attacks have become more common and easier to carry out. A 2017 Cisco Visual Networking Index (VNI) report predicts almost an increase of about two times i.e., 14.5 million of the total number of DDoS attacks by the year 2022. Systems have been proposed for detecting DDoS attacks in the network in an efficient manner. Machine learning is once again being compared to other approaches like intrusion detection systems (IDS), which are commonly utilised for intruder detection and attack type classification. To identify and categorise assaults, the proposed system employs a combination of machine learning methods like xGBoos , KNN, Stochastic Gradient Descent, and Naive Bayes and CNN and deep learning approach. The results demonstrate that XGBoost has the best accuracy, whereas KNN has similar results. The study is strengthened using a hybridization process by employing a bidirectional LSTM (Long Short Term Memory) to obtain more accuracy than the prior method, because the existing method uses machine learning for detection.

K-nearest, Naïve Bayes, Long Short Term Memory (LSTM), Neural Networks, Accuracy