Cross-Validation Results for Intrusion Detection Models on NSL-KDD and CICIDS2017

Cross-validation results from a framework evaluating machine learning models for network intrusion detection. The dataset contains performance metrics from models like Random Forest, which achieved 98.0% accuracy and 97.0% F1-score on harmonized data. The work by Shailendra Mishra was last updated in April 2026.

Use Cases

• Benchmarking supervised and unsupervised learning models based on reported performance metrics.
• Evaluating the impact of feature harmonization on model performance based on the described ablation analysis.
• Comparing model results using statistical tests like Wilcoxon signed-rank and McNemar's tests as mentioned in the framework.
• Assessing reproducibility mechanisms for ML experiments based on the described cryptographic logging approach.

Strengths

• Performance metrics for specific models are provided, including Random Forest with 98.0% accuracy.
• The framework uses two established datasets, NSL-KDD and CICIDS2017, harmonized for comparison.
• Results are statistically validated using multiple tests, as described.
• Dataset is openly licensed under CC-BY-4.0.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The 5.5 KB file size suggests a very limited scope, likely containing summary results rather than raw data.

Provenance

Source

figshare

Collection Method

Results generated from a proposed evaluation framework for IDS systems.

Time Range

null

Freshness

Last updated 2026-04-20 17:36:46; freshness should be verified.

Geography

null