README.md

---
pretty_name: CAD-MiniBooNe
license: cc-by-4.0
configs:
  - config_name: default
    data_files: data.csv
language:
  - en
task_categories:
  - tabular-classification
task_ids:
  - tabular-multi-class-classification
size_categories:
  - 10K<n<100K
tags:
  - anomaly-detection
  - continual-learning
  - continual-anomaly-detection
  - particle-identification
  - physics
  - tabular
---

# CAD-MiniBooNe

## Dataset Summary

**CAD-MiniBooNe** is a single-source continual anomaly detection benchmark scenario derived from the **MiniBooNE Particle Identification** dataset. It recasts the dataset's original binary classification task (distinguishing electron-neutrino "signal" events from muon-neutrino "background" events) as an anomaly detection problem and converts the tabular data into a sequence of concept-grouped tasks.

The dataset contains **66,116 samples**, **5 tasks**, and has a reported **72.47% anomaly ratio in the test set**.


## Intended Use

This dataset is intended for research on:

- continual anomaly detection;
- continual learning for tabular data;
- robustness under distribution shift;
- task ordering in continual-learning benchmarks;
- forgetting and knowledge transfer across related concepts;
- benchmarking anomaly detectors under sequential task exposure.

The intended use is **machine learning research**. As a physics benchmark, it carries no dual-use risk.

## Dataset Source

- MiniBooNE Particle Identification (UCI ML Repository, ID 199): `https://archive.ics.uci.edu/dataset/199/miniboone+particle+identification`

The original dataset originates from the MiniBooNE physics experiment and contains 130,065 events, each described by 50 real-valued particle-identification variables, labeled as electron-neutrino (signal) or muon-neutrino (background) events. Task boundaries in this benchmark are derived via spectral clustering over the original data, used as a proxy for concept drift.


## Dataset Files

The repository contains the following files:

| File | Description |
|---|---|
| `data.csv` | Main tabular dataset file. |
| `orderings.json` | Predefined task orderings for continual-learning evaluation. |
| `croissant.json` | Croissant metadata describing the dataset. |

## Dataset Structure

The main file is:

```text
data.csv
```

The dataset contains task metadata, binary labels, and numerical particle-identification features.

### Core Columns

| Column | Type | Description |
|---|---:|---|
| `task_id` | integer | Numeric identifier of the continual-learning task. |
| `task_name` | string | Name of the task, e.g. `miniboone_0`. |
| `task_split` | string | Split assignment for the row. |
| `label` | integer | Binary anomaly label. `0` denotes normal and `1` denotes anomalous. |

### Task Identifiers

The dataset contains the following task identifiers:

`miniboone_0`, `miniboone_1`, `miniboone_2`, `miniboone_3`, `miniboone_4`

### Feature Columns

The remaining columns are the 50 real-valued particle-identification variables from the original MiniBooNE dataset, provided as `feature_00` … `feature_49`.

For the complete schema, see `croissant.json`.

## Task Orderings

The dataset provides six predefined orderings in `orderings.json`. These orderings define different continual-learning evaluation regimes over the same task set.

| Ordering | Task sequence |
|---|---|
| `curriculum_asc` | `miniboone_2` → `miniboone_3` → `miniboone_0` → `miniboone_4` → `miniboone_1` |
| `curriculum_desc` | `miniboone_1` → `miniboone_4` → `miniboone_0` → `miniboone_3` → `miniboone_2` |
| `generalization_desc` | `miniboone_2` → `miniboone_0` → `miniboone_3` → `miniboone_4` → `miniboone_1` |
| `generalization_asc` | `miniboone_1` → `miniboone_4` → `miniboone_3` → `miniboone_0` → `miniboone_2` |
| `smooth_drift` | `miniboone_1` → `miniboone_4` → `miniboone_0` → `miniboone_3` → `miniboone_2` |
| `abrupt_drift` | `miniboone_0` → `miniboone_1` → `miniboone_2` → `miniboone_4` → `miniboone_3` |

These orderings are intended to expose complementary continual-learning dynamics, including curriculum-like adaptation, generalization-oriented ordering, smooth drift, and abrupt drift.

## Dataset Creation
The details of dataset creation can be found in our paper: [link](https://arxiv.org/abs/2607.18289)

## Citation
When using the dataset, please cite:
```
@article{faber2026towards,
  title={Towards Principled Continual Anomaly Detection: A Systematic Framework and Benchmark Scenarios},
  author={Faber, Kamil and Smendowski, Mateusz and Corizzo, Roberto},
  journal={arXiv preprint arXiv:2607.18289},
  year={2026}
}
```

If you also use the underlying MiniBooNE data, please additionally cite:
```
Roe, B. (2010). MiniBooNE Particle Identification [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5QC87
```