ClarusC64/clinical-oxygen-transport-instability-v0.1

该数据集用于评估模型是否能检测由氧气运输失败引起的不稳定性。每个数据行代表一个简化的氧气输送场景，观察三个时间点。任务目标是判断氧气输送是否保持稳定或趋向不稳定。核心稳定性概念涉及血红蛋白浓度、氧饱和度、心输出量、代谢需求和乳酸积累等因素的相互作用。预测目标为标签1表示氧气运输不稳定，标签0表示稳定。每行数据包括血红蛋白轨迹、氧饱和度轨迹、心输出量代理轨迹、氧气需求代理、乳酸轨迹和干预延迟等变量。此外，还包含干扰变量如实验室噪声和图表噪声。数据集评估指标包括准确率、精确率、召回率、F1分数、混淆矩阵和数据集完整性诊断。

This dataset evaluates whether models can detect instability caused by failure of oxygen transport to tissues. Each row represents a simplified oxygen delivery scenario observed across three time points. The task is to determine whether oxygen delivery remains stable or is moving toward oxygen transport instability. Core stability idea involves interactions between hemoglobin concentration, oxygen saturation, cardiac output, metabolic demand, and lactate accumulation. Prediction target is label = 1 for oxygen transport instability and label = 0 for stable oxygen delivery. Each row includes hemoglobin trajectory, oxygen saturation trajectory, cardiac output proxy trajectory, oxygen demand proxy, lactate trajectory, and intervention delay. Decoy variables include lab_noise and chart_noise. Evaluation metrics include accuracy, precision, recall, f1, confusion matrix, and dataset integrity diagnostics.