gratex/GNOTHEIA-synthetic-insurance-dataset
收藏资源简介:
GNOTHEIA合成保险数据集是一个专门为AI系统设计的合成保险理赔数据集,用于通过OMG SBVR业务规则、结构化理赔多上下文和合成理赔相关文档来评估保险理赔。该数据集的主要目标是支持:LLM微调流程、SBVR推理基准和保险理赔AI评估。数据集完全由GNOTHEIA AI生态系统内的SyDaFa(合成数据工厂)引擎生成,包含863个多上下文记录和490条SBVR规则。多上下文数据集整合了结构化理赔数据、保单数据和理赔文档,而SBVR规则数据集则提供了财产和建筑保险领域的业务规则。数据集覆盖了家庭财产保险(PHI)和建筑保险(PBI)两种产品类型,以及火灾、水损、盗窃等多种风险。所有数据均为合成,不包含真实客户数据或个人可识别信息,适用于AI训练、评估和研究。
The GNOTHEIA Synthetic Insurance Dataset is a synthetic insurance claims dataset designed for AI systems that evaluate insurance claims using OMG SBVR business rules, structured claim polycontexts, and synthetic claim-related documents. The main goals of the dataset are to support: LLM fine-tuning pipeline, SBVR reasoning benchmarks, and insurance claim AI evaluation. The dataset is generated entirely synthetically using the SyDaFa (Synthetic Data Factory) engine within the GNOTHEIA AI ecosystem, containing 863 polycontext records and 490 SBVR rules. The polycontext dataset integrates structured claim data, policy data, and claim documents, while the SBVR rules dataset provides business rules in the field of property and building insurance. The dataset covers two product types: Property Household Insurance (PHI) and Property Building Insurance (PBI), along with various perils such as fire, water damage, and theft. All data is synthetic, with no real customer data or personally identifiable information, making it suitable for AI training, evaluation, and research.





