ClarusC64/clinical-microvascular-instability-v0.1

该数据集用于评估模型是否能检测微血管循环中的不稳定性。每个数据行代表一个简化的微循环监测场景，观察三个时间点。任务目标是确定微血管流动是否保持稳定或趋向于不稳定。数据集包含多个信号指标，如毛细血管流动代理轨迹、氧提取代理轨迹、微血管密度代理轨迹、乳酸轨迹、组织需求代理和干预延迟，以及干扰变量（实验室噪声和图表噪声）。预测目标为二元分类，标签1表示微血管不稳定，标签0表示稳定。数据集还提供了评估方法，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。数据集是Clarus稳定性推理基准的一部分，采用MIT许可证。

This dataset evaluates whether models can detect instability in microvascular circulation. Each row represents a simplified microcirculatory monitoring scenario observed across three time points. The task is to determine whether microvascular flow remains stable or is moving toward microvascular instability. The dataset includes signals such as capillary flow proxy trajectory, oxygen extraction proxy trajectory, microvascular density proxy trajectory, lactate trajectory, tissue demand proxy, and intervention delay, along with decoy variables (lab_noise and chart_noise). The prediction target is binary classification, with label 1 indicating microvascular instability and label 0 indicating stable microcirculation. The dataset also provides evaluation methods, including accuracy, precision, recall, f1, confusion matrix, and dataset integrity diagnostics. The dataset is part of the Clarus Stability Reasoning Benchmark and is licensed under MIT.