|
|
--- |
|
|
license: cc0-1.0 |
|
|
--- |
|
|
# 🧠📚 ArXiv Research Papers Metadata – Finance (q-fin) & Computer Science (cs) |
|
|
|
|
|
This repository provides two extensive and high-quality metadata datasets from [arXiv.org](https://arxiv.org), covering research papers in: |
|
|
|
|
|
- 💸 **Quantitative Finance (q-fin)** |
|
|
- 💻 **Computer Science (cs)** (all subcategories) |
|
|
|
|
|
Each dataset includes structured metadata fields and **direct PDF download links** for every paper. Together, these datasets serve as valuable resources for machine learning research, NLP model training, academic search, and more. |
|
|
|
|
|
--- |
|
|
|
|
|
## 📦 Datasets Included |
|
|
|
|
|
### 1. 🏦 Quantitative Finance (`q-fin`) |
|
|
- 📈 ~Thousands of finance-related papers |
|
|
- ✅ Focused on high-quality papers in areas like: |
|
|
- Financial modeling |
|
|
- Market analysis |
|
|
- Algorithmic trading |
|
|
- Risk management |
|
|
- 🎯 Great for domain-specific LLMs, finance QA, or economic text mining |
|
|
|
|
|
### 2. 💾 Computer Science (`cs`) |
|
|
- 📊 Includes **500,000+** papers across **all CS subfields** |
|
|
- Covers: |
|
|
- Artificial Intelligence (cs.AI) |
|
|
- Machine Learning (cs.LG) |
|
|
- Natural Language Processing (cs.CL) |
|
|
- Computer Vision (cs.CV) |
|
|
- Software Engineering, Theory, Networks, and more |
|
|
- ⚙️ Ideal for LLM pretraining, summarization, or scientific search tools |
|
|
|
|
|
--- |
|
|
|
|
|
## 📑 Features |
|
|
|
|
|
- Direct PDF links to full papers (`pdf_link`) |
|
|
- Rich metadata: |
|
|
- `title`, `authors`, `abstract` |
|
|
- `categories`, `submission date`, `arXiv ID` |
|
|
- Optional: `is_scraped` flag for tracking downloads |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧠 Use Cases |
|
|
|
|
|
- Pretrain or fine-tune LLMs on domain-specific scientific texts |
|
|
- Build citation tools or academic paper search engines |
|
|
- Create offline datasets of research papers |
|
|
- Academic summarization, QA, paper classification, and recommendation systems |
|
|
|
|
|
--- |
|
|
|
|
|
## 📜 License |
|
|
|
|
|
This dataset is released under the **[CC0 1.0 Public Domain Dedication](https://creativecommons.org/publicdomain/zero/1.0/)**. |
|
|
> You are free to copy, modify, distribute, and use this dataset for any purpose — even commercially — without needing permission. |
|
|
|
|
|
--- |
|
|
|
|
|
## 📂 Dataset Format |
|
|
|
|
|
Each row in the CSV files corresponds to a paper. |
|
|
|
|
|
| Column | Description | |
|
|
|----------------|---------------------------------------------------| |
|
|
| `id` | arXiv ID (unique identifier) | |
|
|
| `title` | Paper title | |
|
|
| `authors` | Author names | |
|
|
| `abstract` | Abstract text | |
|
|
| `categories` | arXiv subject classification (e.g., `cs.CL`) | |
|
|
| `pdf_link` | Direct link to the paper’s PDF | |
|
|
| `is_scraped` | (Optional) Indicates if the PDF was downloaded | |
|
|
|
|
|
--- |
|
|
|
|
|
## 📁 Files Included |
|
|
|
|
|
- `qfin_metadata.csv`: Metadata for Quantitative Finance papers |
|
|
- `cs_metadata.csv`: Metadata for Computer Science papers |
|
|
- `qfin_downloader.ipynb`: Script to batch download finance paper PDFs |
|
|
- `cs_downloader.ipynb`: Script to batch download CS paper PDFs |
|
|
|
|
|
--- |
|
|
|
|
|
## 🙋♂️ Maintainer |
|
|
|
|
|
Created and maintained by **[G Gowtham]**. |
|
|
code for gether metadata and pdf's form arXiv avlible at [git hub](https://github.com/Gowthamcopilot/arxiv_papers) |
|
|
Feel free to contribute, ask questions, or report issues! |
|
|
|
|
|
--- |
|
|
|
|
|
## 🙌 Acknowledgements |
|
|
|
|
|
Thanks to [arXiv.org](https://arxiv.org) for making academic research openly accessible. |
|
|
|
|
|
|
