The project is a comparison of 3 different ML classifiers in their prediction accuracy and speed of DDoS attack conditions.
- OneVsRestClassifier
- MultinomialNB
- BaggingClassifier
After using these simple algorithms, they are inturn compared with ensemble algorithms taken a combination of the above 3 mentioned algorithms
- Voting Classifier
- Weighted Classifier
- Stacking Classifier
3 permutations of the weights of the base classifiers are taken for 3 readings of the metrics of the ensemble classifiers
Accuracy(%) Mean Standard Deviation Precision Recall F-Measure True Postive True Negative False Positive
I did not get the permission to upload datasets, five datasets were used each containing network data and DDoS conditions for one kind of protocol The protocols are: ICMP, LAND , TCPSYN, TCPSYNACK , UDP