How bubble-nb/iot-incremental-nids-reproduction works
This repository packages a bounded reproduction and analysis workflow for cross-domain incremental IoT intrusion detection—not the original paper’s full pipeline. The principal workflow loads fixed author-prepared source and target data, trains a source model and target-domain adaptations, evaluates both domains, and saves run artifacts. Label mapping and repeat-run summaries support interpretation; bootstrap scripts prepare the disclosed author-code copies. The sampled excerpts establish the main training and analysis relationships, but do not expose the downloaded framework’s internals.
Main entry points: run_research.py Experiment Runner · bootstrap_author_code.py Bootstrap Code · save_label_mapping.py Label Mapping · summarize_repeat_experiments.py Repeat Summary
Components 15
Experiments
- Experiment Runner
run_research.py - Domain Data
run_research.py - Source Training
run_research.py - Target Adaptations
run_research.py - Domain Evaluation
run_research.py - Run Artifacts
run_research.py
Code Preparation
- Bootstrap Code
bootstrap_author_code.py - Compatibility Copy
prepare_smoke.py - Research Copy
setup_research.py
Result Analysis
- Label Mapping
save_label_mapping.py - Repeat Summary
summarize_repeat_experiments.py - Validation Audit
audit_repeat_validation.py - Reported Results
results/README.md
Other components
- Researcher
- Author Framework
Connections 17
- Researcher → Bootstrap Code: prepares code
README.md - Bootstrap Code → Compatibility Copy: runs preparation
bootstrap_author_code.py - Bootstrap Code → Research Copy: runs preparation
bootstrap_author_code.py - Compatibility Copy → Research Copy: copies and patches
setup_research.py - Researcher → Experiment Runner: starts experiment
README.md - Experiment Runner → Author Framework: imports and reuses
run_research.py - Experiment Runner → Domain Data: loads domains
run_research.py - Experiment Runner → Source Training: trains source model
run_research.py - Experiment Runner → Target Adaptations: trains methods
run_research.py - Experiment Runner → Domain Evaluation: measures domains
run_research.py - Experiment Runner → Run Artifacts: writes outputs
run_research.py - Researcher → Label Mapping: saves mapping
README.md - Label Mapping → Domain Data: loads domains
save_label_mapping.py - Label Mapping → Repeat Summary: runs summary
save_label_mapping.py - Repeat Summary → Run Artifacts: reads predictions
summarize_repeat_experiments.py - Repeat Summary → Reported Results: writes summary
summarize_repeat_experiments.py - Validation Audit → Domain Data: checks splits
audit_repeat_validation.py