Create Readme.md

Dataset Examples/Readme.md

@@ -0,0 +1,35 @@
+## File Descriptions
+
+### `Undersampled UM-NIDS.ipynb`
+This notebook shows how different machine learning models can be trained and evaluated on the standardized flow features provided in the undersampled version of UM-NIDS dataset. 
+### `Multimodal NIDS Comparison.ipynb`
+This notebook compares the performance of NIDS models trained using multimodal features (flow + payload) against models trained using flow features only. It includes an evaluation of the models' robustness and accuracy, showing how multimodal data can improve NIDS performance.
+
+### `Payload cross validation` (folder)
+This folder contains script for cross-validating payload-based NIDS using TF-IDF (Term Frequency-Inverse Document Frequency) representations of payload data. It demonstrates the resilience of payload-based NIDS against payload-specific attacks and compares their performance on various datasets against flow-only models.
+
+### `CIC-IDS 17` (folder)
+This folder contains notebooks for processing the CIC-IDS 2017 dataset. The notebooks include:
+- PCAP file processing and flow feature extraction
+- Model training and evaluation with and without contextual features
+- Handling timestamp issues, timezones, and dataset-specific configurations 
+
+### `CIC-IDS DDoS 19` (folder)
+This folder contains notebooks focused on processing the CIC-IDS DDoS 2019 dataset. It includes:
+- PCAP processing for Distributed Denial of Service (DDoS) attack detection
+- Feature extraction and timestamp alignment
+- Training and evaluating models using flow features with/without contextual data
+
+### `CIC-IoT 23` (folder)
+In this folder, the CIC-IoT 2023 dataset is labeled based on MAC addresses (unlike the other datasets, which use pre-labeled CSV files). The notebooks demonstrate:
+- Custom labeling based on MAC addresses
+- Handling of timestamps and timezones (during labeling)
+- Training models with both flow and payload features for improved performance
+
+### `UNSW` (folder)
+This folder contains processing scripts and evaluation notebooks for the UNSW dataset. The notebooks focus on:
+- PCAP processing and flow feature extraction
+- Handling of timestamps and timezones (during labeling)
+- Handling specific dataset structures and evaluating models with/without contextual features
+
+---