| | --- |
| | language: |
| | - el |
| | tags: |
| | - nlp |
| | - academic |
| | - theses |
| | - opendata |
| | license: cc-by-nc-sa-4.0 |
| | size_categories: |
| | - 10K<n<100K |
| | --- |
| | |
| | **Dataset Moved** |
| | **This dataset is now hosted by the Mozilla Data Collective** |
| |
|
| | Please use the official link below for the most up-to-date version: |
| | **[https://datacollective.mozillafoundation.org/datasets/cmkwvpu7s0032mo07jpk20pj1](https://datacollective.mozillafoundation.org/datasets/cmkwvpu7s0032mo07jpk20pj1)** |
| |
|
| | --- |
| |
|
| | ### Description |
| |
|
| | This dataset is a clean, AI-ready collection of **55,423 doctoral dissertations** (1975–2025) sourced from [OpenArchives.gr](https://www.openarchives.gr) and [didaktorika.gr](https://www.didaktorika.gr). |
| |
|
| | The raw text has been processed using the **glossapi** pipeline—utilizing **DeepSeek OCR** and **Docling**—to ensure accurate capture of polytonic Greek and scientific symbols, standardized into clean **Markdown** for immediate NLP tasks. |
