Hi all,
I want using a dataSet with five features related to three classes ( normal, type attack1, and type attack2 ) to detect the DDoS attack, In this idea which algorithm should be used ( ML with SVM or DL with CNN) and when the SVM will be a better option to using than CNN .. Thanks in advance
Hello,
I'm not sure there's any way to know ahead of time which algorithm will be best for a given problem. Often, experimentation is the only way i.e. you'll need to try both and see which works better.
Informally, some heuristics might be useful. For example, CNNs are most appropriate when your inputs have spatial locality i.e. their spatial positions/arrangement are useful for the task. Based on your problem description, I'd start with the SVM (especially with an RBF kernel), then try a Random Forests classifier and perhaps after that a deep neural network classifier. For five features (which I also suspect to be independent) I'm not sure a CNN would be appropriate, although it wouldn't hurt to try.
Hope this helps.
Good luck!
Dear Tariq Emad Ali ,
If the data you have is not large enough, you are aware of the features it contain, after performing EDA the dataset found not to be highly correlated and instances were linearly seprable, then, SVM is your best choice.
for a detailed comparison of them, you can take a look at https://www.baeldung.com/cs/svm-vs-neural-network.
Regards,
- Badges
- Science topic