Developing the Model

[DRedDevil04]

DDoS Detection Model Challenge

Issue Description

🚀 Challenge: Compete to create the most accurate model for detecting DDoS attacks! Participants are tasked with developing a machine learning model that assigns labels of 1 for DDoS attacks and 0 for normal traffic based on packet features. Additionally, participants are required to maintain a list of source IPs associated with detected DDoS packets.

Evaluation Criteria

• Model Accuracy: The accuracy of the machine learning model in correctly classifying packets.
• False Positive Rate: Minimize false positives to enhance precision.
• List of Source IPs: Maintain an accurate list of source IPs for all packets classified as DDoS attacks.

Data Details

• Three datasets are provided for training and evaluation.
• If you wish to add another dataset, feel free to contribute to the other open-for-all issue and get your data added.
• Features include packet metadata such as IP addresses, TCP/UDP ports, and flags.
• Labels should be binary, with 1 denoting DDoS attacks and 0 denoting normal traffic.

Submission Guidelines

• Participants are encouraged to use a variety of machine learning algorithms and techniques.
• Submissions should include a Jupyter notebook or Python script containing the model implementation.
• Clearly document and comment your code for transparency and understanding.

Reward

• The participant with the most accurate model and well-maintained list of source IPs will be recognized and only their PR will be merged.

Additional Information

• Participants can discuss their approaches, findings, and seek clarifications in the project's discussions.
• Please adhere to project coding standards and guidelines during implementation.

Note

• Refer to /data/CONTRIBUTING.md for information on dataset usage and /scripts/model_evaluation.ipynb for existing evaluation conventions.

[sarthakvermaa]

claim

[DRedDevil04]

@sarthakvermaa it is a competitive issue...no need for claiming it...directly send a PR