README.md

---
language:
- th
license: cc-by-4.0
task_categories:
- audio-classification
task_ids:
- multi-class-classification
tags:
- audio
- thai
- scam
- fraud-detection
- speech
size_categories:
- 10K<n<100K
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: label
    dtype: int64
  splits:
  - name: train
    num_bytes: 12063313
    num_examples: 100
  download_size: 12067276
  dataset_size: 12063313
---

# Thai Scam Call Dataset

## Dataset Description

- **Language:** Thai
- **Task:** Scam Call Audio Classification
- **Total Files:** 21,287 audio clips

This dataset contains **AI-generated** Thai-language telephone audio recordings synthesized to simulate real-world scam and legitimate phone calls. Each sample was generated using text-to-speech (TTS) models with scripted dialogue, designed to closely resemble actual telephone call recordings. The dataset is labeled as **scam (1)** or **not scam (0)** and is intended for training and evaluating audio classification models to detect fraudulent phone calls in Thailand.

Each audio sample is provided in two versions:
- `mp3_15s.zip` — Audio clips trimmed to **15 seconds** (suitable for model training)
- `mp3_full.zip` — Full-length audio recordings

## Labels

| Label | Value | Count | Description |
|-------|-------|-------|-------------|
| not scam | 0 | 11,156 | โทรศัพท์ปกติ (Normal/legitimate call) |
| scam | 1 | 10,131 | โทรศัพท์หลอกลวง (Fraudulent/scam call) |

## Dataset Authors

- Ingkapat J.

## Citation

```
@dataset{thaiscamcall2026,
  author    = {Ingkapat J.},
  title     = {Thai Scam Call Dataset},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/Paam1/ThaiScamCall}
}
```