How geeksgod/ctgan-cicids2017-fedids works
This repository implements a federated intrusion-detection experiment that compares real-only training with CTGAN synthetic-data evaluations (TRTS and TSTR). The documented workflow is to generate synthetic traffic with the CTGAN notebook, provide real and synthetic CSVs, then run the experiment script. The experiment preprocesses data, coordinates organization-level private training and server aggregation, evaluates models, and writes metrics and visualizations. The README documents CTGAN generation, but its notebook implementation was not sampled; its connection to the experiment is therefore documented rather than source-verified.
Main entry points: FederatedIDSExperiment.py Experiment orchestration · imp.py Shared runtime utilities · FederateConfig.py Federated settings · OrganizationAgent.py Organization training
Components 13
Data preparation
- CTGAN generation
CTGAN/600epochCTGAN.ipynb - Data preprocessing
component/DataManager.py
Federated learning
- Experiment orchestration
component/FederatedIDSExperiment.py - IDS neural network
component/CybersecurityNet.py - Organization training
component/OrganizationAgent.py - Federated server
component/FederateServer.py
Evaluation and outputs
- Experiment visualizations
component/Visualizer.py - Experiment results
results
Shared runtime
- Experiment entry
federated_ids_ctgan.py - Federated settings
component/FederateConfig.py - Shared runtime utilities
component/imp.py
Other components
- Researcher
- CICIDS2017 data
Connections 24
- Researcher → CTGAN generation: runs notebook
readme.md - CICIDS2017 data → CTGAN generation: provides dataset
readme.md - CTGAN generation → Experiment orchestration: supplies synthetic CSV
readme.md - Researcher → Experiment entry: runs script
readme.md - Experiment entry → Experiment orchestration: starts experiment
federated_ids_ctgan.py - Experiment entry → Federated settings: sets parameters
federated_ids_ctgan.py - Experiment orchestration → Federated settings: uses settings
component/FederatedIDSExperiment.py - Experiment orchestration → Data preprocessing: loads CSVs
component/FederatedIDSExperiment.py - Data preprocessing → Federated settings: reads feature settings
component/DataManager.py - Experiment orchestration → IDS neural network: creates model
component/FederatedIDSExperiment.py - Experiment orchestration → Federated server: runs federation
component/FederatedIDSExperiment.py - Experiment orchestration → Organization training: creates agents
component/FederatedIDSExperiment.py - Organization training → IDS neural network: trains model
component/OrganizationAgent.py - Organization training → Federated settings: reads training settings
component/OrganizationAgent.py - Federated server → IDS neural network: evaluates global model
component/FederateServer.py - Federated server → Federated settings: reads device settings
component/FederateServer.py - Experiment orchestration → Experiment visualizations: generates plots
component/FederatedIDSExperiment.py - Experiment orchestration → Experiment results: writes metrics
component/FederatedIDSExperiment.py - Experiment entry → Shared runtime utilities: imports utilities
federated_ids_ctgan.py - Experiment orchestration → Shared runtime utilities: imports utilities
component/FederatedIDSExperiment.py - Data preprocessing → Shared runtime utilities: imports utilities
component/DataManager.py - Organization training → Shared runtime utilities: imports utilities
component/OrganizationAgent.py - Federated server → Shared runtime utilities: imports utilities
component/FederateServer.py - Experiment visualizations → Shared runtime utilities: imports utilities
component/Visualizer.py