Code-level experimental result data of deep learning based diagnosis code prediction using patient medication history

Experimental outcomes in terms of accuracy, recall, and F1 for deep-learning-based supervised training of an artificial neural network for patient diagnosis prediction based on patient medication history, using hierarchical domain knowledge to guide learning.

本实验结果以准确率（accuracy）、召回率（recall）与F1值（F1）为评估指标，展示了针对基于患者用药史的患者诊断预测任务，采用层级领域知识指导学习过程的人工神经网络（artificial neural network）所开展的深度学习监督训练的相关结果。