LHRS-UM-FERI/MENTHOS-dataset-spam

---
language:
- en
- sl
tags:
- menthos
- spam
- phishing
- binary-classification
size_categories:
- 10K<n<100K
---

# MENTHOS-dataset-spam

## English

### About

MENTHOS-Spam is a binary text classification dataset for spam/phishing detection. It is created by combining email and SMS spam sources and then producing train/validation/test splits.

### Source Data

Original sources:

- Phishing Email Dataset (Kaggle): https://www.kaggle.com/datasets/naserabdullahalam/phishing-email-dataset
- SMS Spam Collection (Kaggle mirror of UCI dataset): https://www.kaggle.com/datasets/uciml/sms-spam-collection-dataset

### Processing and Balancing

- Unified schema: `text`, `label`
- Label mapping for SMS source: `ham -> 0`, `spam -> 1`
- Split: 70% train, 15% validation, 15% test
- Balancing: each split is balanced by downsampling to the minority class count

### Splits and Class Distribution

The prepared train, validation, and test splits are included with the dataset release.

| split | rows | label 0 | label 1 |
| ---------- | ----: | ------: | ------: |
| train | 60982 | 30491 | 30491 |
| validation | 13166 | 6583 | 6583 |
| test | 13128 | 6564 | 6564 |

### Citation

```
@misc{borovic_li-dobnik_kranjec_ferme_2026,
title = {MENTHOS-dataset-spam},
author = {Borovic, Li Dobnik, Kranjec, Ferme},
year = {2026},
publisher = {Hugging Face},
howpublished = {\url{https://huggingface.co/datasets/LHRS-UM-FERI/MENTHOS-dataset-spam}}
}
```
---

## Slovenščina

### O datasetu

MENTHOS-Spam je binarni dataset za detekcijo spam/phishing besedil. Nastane z združitvijo e-poštnega in SMS vira ter z gradnjo train/validation/test delitev.

### Izvorni podatki

Izvor:

- Phishing Email Dataset (Kaggle): https://www.kaggle.com/datasets/naserabdullahalam/phishing-email-dataset
- SMS Spam Collection (Kaggle ogledalo UCI): https://www.kaggle.com/datasets/uciml/sms-spam-collection-dataset

### Obdelava in uravnoteženje

- Enotna shema: `text`, `label`
- Preslikava oznak za SMS vir: `ham -> 0`, `spam -> 1`
- Delitev: 70% train, 15% validation, 15% test
- Uravnoteženje: vsak split je uravnotežen z downsamplingom na manjšinski razred

### Delitve in porazdelitev razredov

Pripravljene train, validation in test delitve so vključene v izdajo nabora podatkov.

| split | vrstic | label 0 | label 1 |
| ---------- | -----: | ------: | ------: |
| train | 60982 | 30491 | 30491 |
| validation | 13166 | 6583 | 6583 |
| test | 13128 | 6564 | 6564 |

### Citiranje

```
@misc{borovic_li-dobnik_kranjec_ferme_2026,
title = {MENTHOS-dataset-spam},
author = {Borovic, Li Dobnik, Kranjec, Ferme},
year = {2026},
publisher = {Hugging Face},
howpublished = {\url{https://huggingface.co/datasets/LHRS-UM-FERI/MENTHOS-dataset-spam}}
}
```