IRBXrocket/phishing-url

---
license: cc-by-4.0
configs:
- config_name: default
data_files:
- split: train
path: data/train.parquet
- split: test
path: data/test.parquet
task_categories:
- text-classification
- tabular-classification
- token-classification
- text2text-generation
size_categories:
- n<1K
annotations_creators:
- found
tags:
- phishing
- url
- security
language:
- en
pretty_name: TabNetone
---

# Dataset Description

The provided dataset includes **11430** URLs with **87** extracted features.
The dataset are designed to be used as a benchmark for machine learning based **phishing detection** systems.
The datatset is balanced, it containes exactly 50% phishing and 50% legitimate URLs.

Features are from three different classes:
- **56** extracted from the structure and syntax of URLs
- **24** extracted from the content of their correspondent pages
- **7** are extracetd by querying external services.

The dataset was partitioned randomly into training and testing sets, with a ratio of **two-thirds for training** and **one-third for testing**.

## Details

- **Funded by:** Abdelhakim Hannousse, Salima Yahiouche
- **Shared by:** [pirocheto](https://github.com/pirocheto)
- **License:** [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)
- **Paper:** [https://arxiv.org/abs/2010.12847](https://arxiv.org/abs/2010.12847)

## Source Data

The diagram below illustrates the procedure for creating the corpus.
For details, please refer to the paper.

<div align="center">
<img src="images/source_data.png" alt="Diagram source data">
</div>

<p align="center">
<em>Source: Extract form the <a href="https://arxiv.org/abs/2010.12847">paper</a></em>
</p>

## Load Dataset

- With **datasets**:
```python
from datasets import load_dataset

dataset = load_dataset("pirocheto/phishing-url")
```

- With **pandas** and **huggingface_hub**:
```python
import pandas as pd
from huggingface_hub import hf_hub_download

REPO_ID = "pirocheto/phishing-url"
FILENAME = "data/train.parquet"

df = pd.read_parquet(
hf_hub_download(repo_id=REPO_ID, filename=FILENAME, repo_type="dataset")
)
```

- With **pandas** only:
```python
import pandas as pd

url = "https://huggingface.co/datasets/pirocheto/phishing-url/resolve/main/data/train.parquet"
df = pd.read_parquet(url)
```

## Citation

To give credit to the creators of this dataset, please use the following citation in your work:

- BibTeX format

```
@article{Hannousse_2021,
title={Towards benchmark datasets for machine learning based website phishing detection: An experimental study},
volume={104},
ISSN={0952-1976},
url={http://dx.doi.org/10.1016/j.engappai.2021.104347},
DOI={10.1016/j.engappai.2021.104347},
journal={Engineering Applications of Artificial Intelligence},
publisher={Elsevier BV},
author={Hannousse, Abdelhakim and Yahiouche, Salima},
year={2021},
month=sep, pages={104347} }
```

- APA format

```
Hannousse, A., & Yahiouche, S. (2021).
Towards benchmark datasets for machine learning based website phishing detection: An experimental study.
Engineering Applications of Artificial Intelligence, 104, 104347.
```

---
许可证：CC-BY-4.0
配置项：
- 配置名称：default
数据文件：
- 拆分集：训练集（train），路径：data/train.parquet
- 拆分集：测试集（test），路径：data/test.parquet
任务类别：
- 文本分类（text-classification）
- 表格分类（tabular-classification）
- Token分类（token-classification）
- 文本到文本生成（text2text-generation）
样本规模类别：n<1K
标注创作者：公开采集标注（found）
标签：
- 钓鱼（phishing）
- URL
- 安全（security）
语言：英语（en）
展示名称：TabNetone
---

# 数据集描述

本数据集包含11430条统一资源定位符（URL），并提取了87项特征。本数据集旨在作为基于机器学习的钓鱼检测（phishing detection）系统的基准测试数据集。本数据集为平衡数据集，其中钓鱼URL与合法URL各占50%。

特征分为三大类：
- **56**项特征提取自URL的结构与语法
- **24**项特征提取自对应网页的内容
- **7**项特征通过调用外部服务提取。

本数据集已按随机方式划分为训练集与测试集，划分比例为训练集占三分之二，测试集占三分之一。

## 详细信息

- **资助方：** Abdelhakim Hannousse、Salima Yahiouche
- **共享方：** [pirocheto](https://github.com/pirocheto)
- **许可证：** [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)
- **相关论文：** [https://arxiv.org/abs/2010.12847](https://arxiv.org/abs/2010.12847)

## 源数据构建流程

下图展示了该数据集语料库的构建流程，详细流程请参阅上述论文。

<div align="center">
<img src="images/source_data.png" alt="源数据示意图">
</div>
<p align="center">
<em>来源：摘自<a href="https://arxiv.org/abs/2010.12847">论文</a></em>
</p>

## 数据集加载方式

- 使用**datasets**库加载：
python
from datasets import load_dataset

dataset = load_dataset("pirocheto/phishing-url")

- 使用**pandas**与**huggingface_hub**库加载：
python
import pandas as pd
from huggingface_hub import hf_hub_download

REPO_ID = "pirocheto/phishing-url"
FILENAME = "data/train.parquet"

df = pd.read_parquet(
hf_hub_download(repo_id=REPO_ID, filename=FILENAME, repo_type="dataset")
)

- 仅使用**pandas**库加载：
python
import pandas as pd

url = "https://huggingface.co/datasets/pirocheto/phishing-url/resolve/main/data/train.parquet"
df = pd.read_parquet(url)

## 引用规范

若在研究工作中使用本数据集，请采用以下引用格式以标注原作者：

- BibTeX引用格式：

@article{Hannousse_2021,
title={Towards benchmark datasets for machine learning based website phishing detection: An experimental study},
volume={104},
ISSN={0952-1976},
url={http://dx.doi.org/10.1016/j.engappai.2021.104347},
DOI={10.1016/j.engappai.2021.104347},
journal={Engineering Applications of Artificial Intelligence},
publisher={Elsevier BV},
author={Hannousse, Abdelhakim and Yahiouche, Salima},
year={2021},
month=sep, pages={104347} }

- APA引用格式：

Hannousse, A., & Yahiouche, S. (2021).
Towards benchmark datasets for machine learning based website phishing detection: An experimental study.
Engineering Applications of Artificial Intelligence, 104, 104347.