Health Insurance Decision-Making Dataset
收藏资源简介:
该数据集模拟了个人在不确定的情况下选择健康保险计划时所面临的困境,包含来自真实计划规范的20个成本相关的探测问题,每个问题都与整数美元金额的替代计划选项配对。决策者行为由两个认知属性描述:风险容忍度和选择,每个属性都标记为高或低,产生四种可能的组合。数据集还包括家庭组成、医疗历史、就业类型和生活方式因素等上下文特征,以描述场景。每个实例包括四个计划选项,一个标记的属性和与目标决策者对齐的真实选择。数据集旨在训练和评估算法决策者,用于决策者对齐研究。
This dataset simulates the dilemmas encountered by individuals when selecting health insurance plans under uncertainty. It contains 20 cost-related probing questions sourced from real health insurance plan specifications, with each question paired against alternative plan options denominated in whole US dollar amounts. Decision-maker behavior is defined by two cognitive attributes: risk tolerance and choice preference, each categorized as either high or low, yielding four distinct combinations. The dataset also incorporates contextual features such as household composition, medical history, employment type, and lifestyle factors to fully describe the decision-making scenarios. Each instance includes four plan options, one labeled attribute, and the ground-truth choice aligned with the target decision-maker. This dataset is designed to train and evaluate algorithmic decision-makers for research on decision-maker alignment.




