From: Detecting Denial of Service attacks using machine learning algorithms
Authors | Their contributions | Proposed work |
---|---|---|
Zekri M et.al [1] | Usage of C.4.5 algorithm and Signature detection algorithms were developed for automatically and successfully detecting signature attacks for DDoS attacks | Machine learning technologies such as logistic regression and naive Bayes are also used to detect intrusion |
Xiaoyong Yuan et. Al [2] | A DDoS assault detection strategy based on deep learning was proposed and then compared to a regular machine learning model | To detect DDoS attacks, a mathematical model and a machine learning model were suggested, and the miss rate for each was then compared |
Ceron J et.al [23] | Botnet assaults were detected using the UNBS-NB 15 and KDD99 publicity datasets | Botnet assaults were detected using CAIDA 2007 records |
Mohamed Idhammad et.al [13] | An interactive, progressive, semi-supervised ML algorithm for DDoS attack detection was created using connectivity diversity prediction, co-clustering, mutual information proportion, and the Extra-Trees methodology | To identify patterns, the dataset was partitioned into 70:20:10 train, test, and validation categories, and ML strategies, notably logistic regression and naive bayes, were implemented |
Saikat Das et.al [21] | NSL-KDD dataset was tested and analysed | CAIDA 2007 dataset was tested and analysed |
Qian Li et.al [22] | For the real dataset, PCA-RNN was examined and shown to have significant performance increases in terms of accuracy, sensitivity, precision, and F-score | On a real-world dataset, naive bayes and logistic regression were reported to have positive performance improvements in terms of accuracy, sensitivity, precision, and F-score |
Tong Anh Tuan et.al [24] | Support Vector Machine (SVM), Artificial Neural Network (ANN), Naive Bayes (NB), Decision Tree (DT), and semi-supervised learning have been used to evaluate and compare the UNBS-NB 15 dataset (USML) | The CAIDA 2007 dataset was analysed and compared to logistic regression, naive bayes, and experimented findings, with the results for both being summarised together |