Network Intrusion Classification Data Split Strategy

Khorshed Alam published a dataset on figshare in April 2026 detailing a training-testing split strategy for network intrusion classification. The work proposes a deep reinforcement learning approach integrated with Echo State Networks and evaluates data balancing techniques like Borderline-SMOTE and K-means SMOTE. Validation was conducted on benchmark datasets including NF-BoT-IoT and NF-CSE-CIC-IDS2018.

Use Cases

• Testing adaptive intrusion detection models based on excluded attack types like DoS and Backdoor
• Evaluating data balancing techniques for network traffic data based on methods like K-means SMOTE
• Benchmarking intrusion detection algorithms across multiple network flow datasets mentioned in the description
• Training reinforcement learning agents for real-time threat detection based on the DRL approach described

Strengths

• Multi-dataset validation on six benchmark datasets ensures robustness across different network flows
• The dataset is 5.5 KB and available in XLS format, facilitating easy access and inspection
• Source code is provided in a linked GitHub repository for reproducibility

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• The description focuses on methodology; the actual data content and structure are not detailed

Provenance

Source

figshare

Collection Method

Likely derived from benchmark intrusion detection datasets and processed for a specific research methodology.

Freshness

Last updated 2026-04-16 17:40:22