Cyberattacks are becoming more complex, it is getting harder to detect specific intrusions. The techniques used here for detection are based on machine learning techniques, The technique of extracting useful knowledge from huge amounts of data.
Project steps:
-
Data analysis and EDA
- Number of instances per class
- Number of possible outliers
- Number of null values per feature
- Feaure Engineering
- One Hot Encoding
- Label Encoding
- Scaling
- Modelling
-
MLP – Base line
-
-
Optimaization
- change the number of hidden layes
- change number of nuerons
- change batch size
- change the learning rate
- change the optimizer
- change the activation function