CODE-15%

该数据集是我们用于训练模型的最大数据集，它经过特别筛选，以确保模型的泛化能力，同时排除了来自PTB的数据。该数据集被用来抽取正负样本对进行训练，确保了训练过程中的平衡表现。由于数据规模过大，不适宜进行枚举，因此我们对其进行抽样以用于训练和验证。这项任务旨在训练一个神经网络模型，用于患者识别以及检测记录分配错误。

This dataset is the largest one employed for our model training. It was subjected to specialized screening to guarantee the model’s generalization capability, while excluding all data sourced from PTB. Positive and negative sample pairs were extracted from this dataset for training, ensuring balanced performance during the training process. Given its excessively large scale which makes enumeration infeasible, we performed sampling on the dataset for training and validation purposes. The objective of this task is to train a neural network model for patient identification and the detection of record allocation errors.

搜集汇总

背景与挑战

背景概述

CODE-15%是一个大规模标注的12导联心电图数据集，包含来自23万患者的34万份检查数据，具有详细的临床标注和原始信号数据。该数据集是CODE数据集的代表性子集，主要用于心电图自动诊断和死亡率预测研究。