README.md

metadata

task_categories:
  - text-classification
  - token-classification
language:
  - en
tags:
  - knowledge-graph
  - relation-extraction
  - synthetic-data
size_categories:
  - 1K<n<10K

Synthetic Knowledge Graph Dataset

Dataset Description

This is a synthetic dataset for knowledge graph relation extraction, generated from a scale-free knowledge graph.

Dataset Summary

• Total samples: 160
• Total triples: 160
• Relations: 14

Supported Tasks

• Relation extraction
• Knowledge graph construction
• Natural language understanding

Dataset Structure

Each sample contains:

• sentence: Natural language sentence
• subject: Subject entity
• relation: Relation type
• object: Object entity
• triple: Formatted triple

Data Fields

{
  "sentence": "Alice is a friend of Bob.",
  "subject": "Alice", 
  "relation": "friend_of",
  "object": "Bob",
  "subject_type": "name",
  "object_type": "name", 
  "triple": "(Alice, friend_of, Bob)",
  "template_id": 0
}

Usage

from datasets import load_dataset

dataset = load_dataset("your-username/synthetic-kg-dataset")