ClarusC64/clinical-organ-coupling-cascade-v0.1

该数据集用于评估模型是否能检测到由多器官耦合引起的不稳定性。每一行数据代表心血管、代谢、呼吸、肾脏和血液学指标的短轨迹。数据集的核心思想是临床崩溃经常发生在器官间的压力信号相互加强时，例如氧合下降与乳酸升高相结合，肾脏压力与代谢恶化相结合，血小板下降与循环不稳定相结合。预测目标是标签1表示由于多器官级联引起的不稳定性，标签0表示稳定的多器官轨迹。每行数据包含平均动脉压轨迹、乳酸轨迹、肌酐轨迹、氧饱和度轨迹、血小板轨迹、液体反应指标和通气支持标志，以及一些干扰变量。数据集是ClarusC64稳定性推理基准系列的一部分，支持跨领域稳定性推理和多变量级联检测的研究。

This dataset evaluates whether models can detect instability arising from multi-organ coupling. Each row represents a short trajectory across cardiovascular, metabolic, respiratory, renal, and hematologic indicators. The core stability idea is that clinical collapse frequently occurs when stress signals across organs reinforce each other, such as declining oxygenation combined with rising lactate, renal stress combined with metabolic deterioration, and platelet decline combined with circulatory instability. The prediction target is label = 1 for instability due to multi-organ cascade and label = 0 for stable multi-organ trajectory. Each scenario contains MAP trajectory, lactate trajectory, creatinine trajectory, oxygen saturation trajectory, platelet trajectory, fluid response indicator, ventilation support flag, and some decoy variables. This dataset is part of the ClarusC64 stability-reasoning benchmark family and supports research into cross-domain stability reasoning and multi-variable cascade detection.