--- dataset_info: features: - name: flow_id dtype: string - name: features struct: - name: src_port dtype: float32 - name: dst_port dtype: float32 - name: protocol dtype: float32 - name: flow_duration dtype: float32 - name: total_fwd_packet dtype: float32 - name: total_bwd_packets dtype: float32 - name: total_length_of_fwd_packet dtype: float32 - name: total_length_of_bwd_packet dtype: float32 - name: fwd_packet_length_max dtype: float32 - name: fwd_packet_length_min dtype: float32 - name: fwd_packet_length_mean dtype: float32 - name: fwd_packet_length_std dtype: float32 - name: bwd_packet_length_max dtype: float32 - name: bwd_packet_length_min dtype: float32 - name: bwd_packet_length_mean dtype: float32 - name: bwd_packet_length_std dtype: float32 - name: flow_bytes_per_s dtype: float32 - name: flow_packets_per_s dtype: float32 - name: flow_iat_mean dtype: float32 - name: flow_iat_std dtype: float32 - name: flow_iat_max dtype: float32 - name: flow_iat_min dtype: float32 - name: fwd_iat_total dtype: float32 - name: fwd_iat_mean dtype: float32 - name: fwd_iat_std dtype: float32 - name: fwd_iat_max dtype: float32 - name: fwd_iat_min dtype: float32 - name: bwd_iat_total dtype: float32 - name: bwd_iat_mean dtype: float32 - name: bwd_iat_std dtype: float32 - name: bwd_iat_max dtype: float32 - name: bwd_iat_min dtype: float32 - name: fwd_psh_flags dtype: float32 - name: bwd_psh_flags dtype: float32 - name: fwd_urg_flags dtype: float32 - name: bwd_urg_flags dtype: float32 - name: fwd_header_length dtype: float32 - name: bwd_header_length dtype: float32 - name: fwd_packets_per_s dtype: float32 - name: bwd_packets_per_s dtype: float32 - name: packet_length_min dtype: float32 - name: packet_length_max dtype: float32 - name: packet_length_mean dtype: float32 - name: packet_length_std dtype: float32 - name: packet_length_variance dtype: float32 - name: fin_flag_count dtype: float32 - name: syn_flag_count dtype: float32 - name: rst_flag_count dtype: float32 - name: psh_flag_count dtype: float32 - name: ack_flag_count dtype: float32 - name: urg_flag_count dtype: float32 - name: cwr_flag_count dtype: float32 - name: ece_flag_count dtype: float32 - name: down_per_up_ratio dtype: float32 - name: average_packet_size dtype: float32 - name: fwd_segment_size_avg dtype: float32 - name: bwd_segment_size_avg dtype: float32 - name: fwd_bytes_per_bulk_avg dtype: float32 - name: fwd_packet_per_bulk_avg dtype: float32 - name: fwd_bulk_rate_avg dtype: float32 - name: bwd_bytes_per_bulk_avg dtype: float32 - name: bwd_packet_per_bulk_avg dtype: float32 - name: bwd_bulk_rate_avg dtype: float32 - name: subflow_fwd_packets dtype: float32 - name: subflow_fwd_bytes dtype: float32 - name: subflow_bwd_packets dtype: float32 - name: subflow_bwd_bytes dtype: float32 - name: fwd_init_win_bytes dtype: float32 - name: bwd_init_win_bytes dtype: float32 - name: fwd_act_data_pkts dtype: float32 - name: fwd_seg_size_min dtype: float32 - name: active_mean dtype: float32 - name: active_std dtype: float32 - name: active_max dtype: float32 - name: active_min dtype: float32 - name: idle_mean dtype: float32 - name: idle_std dtype: float32 - name: idle_max dtype: float32 - name: idle_min dtype: float32 - name: semantic_flags struct: - name: high_packet_rate dtype: int8 - name: one_way_traffic dtype: int8 - name: syn_ack_imbalance dtype: int8 - name: uniform_packet_size dtype: int8 - name: long_idle_c2 dtype: int8 - name: label dtype: class_label: names: '0': benign '1': analysis '2': backdoor '3': dos '4': exploits '5': fuzzers '6': generic '7': reconnaissance '8': shellcode '9': worms - name: is_attack dtype: int8 splits: - name: train num_bytes: 248798843 num_examples: 671088 - name: validation num_bytes: 62199963 num_examples: 167772 - name: test num_bytes: 77749889 num_examples: 209715 download_size: 119275343 dataset_size: 388748695 configs: - config_name: default data_files: - split: train path: data/train-* - split: validation path: data/validation-* - split: test path: data/test-* task_categories: - tabular-classification language: - en tags: - cybersecurity size_categories: - 1M *Given a network flow represented by CICFlowMeter features, predict the attack category.* --- ## Label Taxonomy The dataset uses the following **10-class label space**: | Label ID | Name | Description | | -------: | -------------- | ------------------------------------------ | | 0 | benign | Normal network traffic | | 1 | analysis | Port scans, probing, and analysis activity | | 2 | backdoor | Backdoor and remote access behavior | | 3 | dos | Denial-of-Service attacks | | 4 | exploits | Exploitation of vulnerabilities | | 5 | fuzzers | Fuzzing and malformed input attacks | | 6 | generic | Generic attack traffic | | 7 | reconnaissance | Reconnaissance and information gathering | | 8 | shellcode | Shellcode execution attempts | | 9 | worms | Worm propagation traffic | ### Additional label fields * `label` → multiclass label (0–9) * `is_attack` → binary indicator (`1` if label ≠ 0, else `0`) This allows both **multiclass IDS** and **binary detection** experiments without reprocessing. --- ## Dataset Structure ```json DatasetDict({ train, validation, test }) ``` Each split contains records with the following schema: ```json flow_id: string features: dict[str, float] semantic_flags: dict[str, int] label: ClassLabel (0–9) is_attack: int (0/1) ``` --- ## Feature Description ### 1. Raw Numeric Features (`features`) The `features` field contains **CICFlowMeter-derived flow statistics**, including: #### Traffic volume & direction * `total_fwd_packets`, `total_bwd_packets` * `total_length_of_fwd_packets` * `total_length_of_bwd_packets` * `down_up_ratio` #### Packet size statistics * `packet_length_min`, `packet_length_max` * `packet_length_mean`, `packet_length_std`, `packet_length_variance` * Forward and backward packet length statistics #### Timing & inter-arrival times * `flow_duration` * `flow_iat_mean`, `flow_iat_std`, `flow_iat_max`, `flow_iat_min` * Forward and backward IAT statistics * `active_*`, `idle_*` metrics #### Rate-based features * `flow_bytes_per_s` * `flow_packets_per_s` * `fwd_packets_per_s`, `bwd_packets_per_s` #### TCP flag counters * `syn_flag_count`, `ack_flag_count`, `rst_flag_count` * `fin_flag_count`, `psh_flag_count`, `urg_flag_count` * `cwr_flag_count`, `ece_flag_count` #### Bulk and subflow statistics * `*_bulk_*` features (conditionally emitted by CICFlowMeter) * `subflow_fwd_*`, `subflow_bwd_*` > **Note:** Some CICFlowMeter features are conditionally emitted. > Missing or undefined numeric values were **filled with `0.0`**, which semantically indicates *absence of that behavior*. --- ### 2. Semantic Flags (`semantic_flags`) To support **interpretable and rule-based IDS**, each flow includes deterministic semantic indicators: | Flag | Meaning | | --------------------- | ---------------------------------------- | | `high_packet_rate` | Very high packets per second | | `one_way_traffic` | Forward-only traffic (no response) | | `syn_ack_imbalance` | Large SYN/ACK imbalance | | `uniform_packet_size` | Low packet size variance | | `long_idle_c2` | Long idle periods (possible C2 behavior) | These flags are: * Derived deterministically from numeric features * Stable across dataset versions * Suitable for rule-based or hybrid ML systems --- ## Preprocessing Summary The dataset was produced using a **fully deterministic pipeline**: * Column name normalization * Conversion of `NaN` / `±Infinity` → `0.0` * Explicit label normalization (string → numeric) * No row drops based on missing numeric values * Stratified train/validation/test splits * Explicit Hugging Face feature schemas No synthetic balancing, oversampling, or augmentation was applied. --- ## Class Distribution & Imbalance The dataset is **highly imbalanced**, reflecting real-world network traffic: * Benign traffic dominates * Some attack classes are rare This is **intentional and realistic**. Users are strongly encouraged to: * Use **macro / weighted F1** * Inspect **per-class recall** * Apply **class weighting** where appropriate Accuracy alone is not a meaningful metric for this dataset. --- ## Related Datasets * A **binary IDS version** of this dataset is published separately for attack detection use cases. * Both datasets share identical features and preprocessing logic. ---