Add paper link, GitHub link, task category, and sample usage

README.md

@@ -1,5 +1,7 @@
 ---
 license: cc-by-4.0
+task_categories:
+- text-classification
 dataset_version: 0.2.3
 dataset_info:
   features:
@@ -243,10 +245,43 @@ configs:
   - split: test
     path: data/test-*
 ---
-# ds4dh/ct-dosing-errors
+
+# Dataset Card for ct-dosing-errors
+
+This dataset provides the materials accompanying the paper "[Early Risk Stratification of Dosing Errors in Clinical Trials Using Machine Learning](https://huggingface.co/papers/2602.22285)".
+
+The dataset is designed for the prediction of dosing errors in interventional clinical research. It comprises 42,112 clinical trials extracted from ClinicalTrials.gov, containing structured, semi-structured trial data, and unstructured protocol-related free-text data.
+
+## Links
+
+- **Paper:** [https://huggingface.co/papers/2602.22285](https://huggingface.co/papers/2602.22285)
+- **GitHub Repository:** [https://github.com/ds4dh/CT-dosing-errors](https://github.com/ds4dh/CT-dosing-errors)
+
+## Sample Usage
+
+You can load the dataset using the `datasets` library:
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset(
+    "ds4dh/ct-dosing-errors",
+    split="train"
+)
+
+print(ds)
+print(ds.features)
+```
+
+## Dataset Summary
+
+The dataset includes labels for the prediction of dosing errors derived from adverse event reports and MedDRA terminology. It features a wide range of fields including:
+- **Textual data:** Brief summaries, detailed descriptions, conditions, and protocol PDF text.
+- **Structured data:** Clinical trial phases, enrollment counts, allocation, and intervention types.
+- **Labels:** Binary indicators (`LABEL_wilson_label`) for elevated dosing error rates.
 
 ## Version
 This repository contains dataset version **0.2.3**.
 
 ## License
-This dataset is licensed under **CC BY 4.0** (`cc-by-4.0`).
+This dataset is licensed under **CC BY 4.0** (`cc-by-4.0`).