asbatel / containerhids Goto Github PK

From testing and reading the source code, I don't find it very clear which files shall be in the so-called training folder. I understand it shall contain traces,seen syscalls, seen args and the max seq freq. However, the format of those and their filename are specified nowhere.
If I would like to train the model on the cb-ds dataset for example, how do I proceed?

Thanks in advance, it will save me to do some reverse eng! :)

Thank you for your great work. I'm wondering which version of LID-DS you used. I noticed that LID-DS 2019 has a .csv file to tell whether a record is normal and LID-DS 2021 separates the dataset into training, testing and validation. I suppose that if I use the 2021 dataset I shall just use the records in the training folder as it contains only normal data, and if I use the 2019 dataset I would need to manually pick out the normal records for training.
Another question is that I find in the CB-DS dataset 1000 scap files for normal record and 100 or 101 scap files for each attack, while the paper says 1700 for normal and 99 for each attack. I wonder what made the difference.
I'd appreciate it if you could answer my confusion😊.