Enterprise SQL Schema in the Insurance Domain

本数据集名为‘企业保险领域SQL架构’，由data.world机构创建，包含43个自然语言问题，涵盖从简单到复杂的业务报告和关键绩效指标问题。数据集基于OMG财产和意外伤害数据模型，是一个企业关系数据库架构。此外，数据集还包括一个上下文层，包含描述保险领域的业务概念、属性和关系的本体，以及从SQL架构到本体的映射，用于创建知识图谱。该数据集旨在评估大型语言模型在企业环境中对SQL数据库进行问答的准确性，并探讨知识图谱在提高准确性方面的作用。

This dataset, named "Enterprise Insurance Domain SQL Schema", was created by data.world. It contains 43 natural language questions covering simple to complex business reporting and key performance indicator (KPI)-related queries. Built on the OMG Property and Casualty Data Model, the dataset is an enterprise-grade relational database schema. Additionally, the dataset includes a contextual layer that comprises an ontology describing business concepts, attributes and relationships within the insurance domain, as well as mappings from the SQL schema to the ontology for knowledge graph construction. This dataset is designed to evaluate the accuracy of large language models (LLMs) performing question answering over SQL databases in enterprise environments, and to explore the role of knowledge graphs in enhancing such accuracy.