Dataset: ArabSummaries

About this Dataset

ArabSummaries

ArabSummaries is a curated dataset designed for Arabic text summarization tasks. It consists of 7,000 Arabic documents spanning various domains, paired with their corresponding summaries generated using advanced AI models. The dataset provides a rich resource for natural language processing (NLP) research, especially in the field of text summarization.

Key Features:
- Domain Coverage:
1,000 documents per domain, covering Culture, Medical, Politics, Religion, Sports, and Technology.

- Data Source:
The original documents are derived from the SANAD Dataset.https://www.kaggle.com/datasets/haithemhermessi/sanad-dataset/data

- Summaries:
Generated using state-of-the-art AI summarization models, ensuring high-quality, concise representations of the original texts.

Potential Use Cases:

Training and evaluation of Arabic text summarization models.
Domain-specific summarization research and analysis.
Fine-tuning and benchmarking NLP models for Arabic-language applications.

ArabSummaries bridges the gap in Arabic summarization resources, providing a diverse and comprehensive dataset for researchers and developers.

Tables

Arabsummaries

@kaggle.tahaalselwii_arabsummaries.arabsummaries

15.29 MB
7000 rows
5 columns


CREATE TABLE arabsummaries (
  "text" VARCHAR,
  "summary" VARCHAR,
  "type" VARCHAR,
  "text_length" BIGINT,
  "summary_length" BIGINT
);