A dataset of knowledge injection for Gansu tourism large models

Through data collection from multiple tourism platforms and manual curation, the dataset has constructed a knowledge graph containing entities and relationships, totaling 810,793 triples, involving 30,045 entities and 14 types of relationships.

The dataset is stored in CSV format and uses Neo4j for storage and visualization, facilitating intuitive data exploration and analysis. In addition to providing basic information about tourism entities, the dataset also offers detailed descriptions of entities through dimensions such as geographic location, ratings, and recommended dishes. This dataset is suitable for applications such as tourism recommendation systems and intelligent question answering, providing rich training materials and knowledge support for tourism systems based on large models, and contributing to the enhancement of the system's knowledge coverage and reasoning abilities.