The goal of this project is to build a machine learning model that is able to detect and classify network attacks. The model will be trained using a dataset of network flows, and will be able to predict if a new flow is an attack or not, and if it is, what kind of attack it is.
The dataset used for this project is written in the NetFlow V9 format (format by Cisco, documentation available here). The dataset is composed by two files:
- train_set: ~4 million flows, used for training the model
- test_set: ~2 million flows, used for testing the model