ClarusC64/clinical-glucose-regulation-instability-v0.1

该数据集用于评估模型是否能检测葡萄糖调节的不稳定性。每个数据行代表一个简化的葡萄糖控制场景，观察三个时间点。任务目标是确定代谢葡萄糖调节是否保持稳定或趋向于调节不稳定。数据集包含葡萄糖轨迹、胰岛素反应代理轨迹、肝缓冲代理轨迹、酮体代理轨迹、代谢需求代理和干预延迟等变量。预测目标为label=1表示葡萄糖调节不稳定，label=0表示稳定的代谢葡萄糖控制。数据集还包含干扰变量如实验室噪声和图表噪声。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。数据集是Clarus稳定性推理基准的一部分。

This dataset evaluates whether models can detect instability in glucose regulation. Each row represents a simplified glucose control scenario observed across three time points. The task is to determine whether metabolic glucose regulation remains stable or is moving toward regulatory instability. The dataset includes variables such as glucose trajectory, insulin response proxy trajectory, hepatic buffer proxy trajectory, ketone proxy trajectory, metabolic demand proxy, and intervention delay. The prediction target is label=1 for glucose regulation instability and label=0 for stable metabolic glucose control. The dataset also includes decoy variables like lab_noise and chart_noise. Evaluation metrics include accuracy, precision, recall, f1, confusion matrix, and dataset integrity diagnostics. This dataset is part of the Clarus Stability Reasoning Benchmark.