Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
json
Languages:
English
Size:
1K - 10K
Tags:
text-classification
scientific-posters
poster-detection
poster-sentry
machine-actionable
FAIR-data
License:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- text-classification
|
| 7 |
+
- scientific-posters
|
| 8 |
+
- poster-detection
|
| 9 |
+
- poster-sentry
|
| 10 |
+
size_categories:
|
| 11 |
+
- 1K<n<10K
|
| 12 |
+
task_categories:
|
| 13 |
+
- text-classification
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# PosterSentry Training Data
|
| 17 |
+
|
| 18 |
+
Extracted text from real scientific poster PDFs and non-poster documents,
|
| 19 |
+
used to train the [PosterSentry classifier](https://huggingface.co/fairdataihub/poster-sentry).
|
| 20 |
+
|
| 21 |
+
## Format
|
| 22 |
+
|
| 23 |
+
NDJSON with `text` and `label` fields:
|
| 24 |
+
```json
|
| 25 |
+
{"text": "TITLE: Effects of Temperature on...", "label": "poster"}
|
| 26 |
+
{"text": "Abstract. We present a novel approach...", "label": "non_poster"}
|
| 27 |
+
```
|
| 28 |
+
|
| 29 |
+
## Statistics
|
| 30 |
+
|
| 31 |
+
| Label | Count | Source |
|
| 32 |
+
|-------|-------|--------|
|
| 33 |
+
| poster | 1,803 | Zenodo & Figshare verified posters |
|
| 34 |
+
| non_poster | 1,803 | Multi-page docs (papers, proceedings, newsletters) |
|
| 35 |
+
|
| 36 |
+
## Source Corpus
|
| 37 |
+
|
| 38 |
+
Sampled from 30K+ PDFs in the [posters.science](https://posters.science) collection:
|
| 39 |
+
- 28,111 verified scientific posters
|
| 40 |
+
- 2,036 verified non-posters
|