基于深度神经网络的儿科临床辅助决策算法

Pediatric Clinical Decision Support Algorithm Based on Deep Neural Network

目的:使用深度神经网络对儿科电子病历进行挖掘分析,实现临床辅助决策支持。方法:首先,对非结构化电子病历数据进行预处理,然后利用自然语言处理技术将其转化为句向量。其次,构建双向循环神经网络(BiRNN)模型,用来捕捉患者的临床症状与多重环境因素之间的复杂关联关系。最后,面向149 817条儿科电子病历数据集来训练和验证模型。结果:实验结果表明,提出的基于双向循环神经网络的儿科临床辅助决策算法的预测精度优于四种对比算法。

Objective: We use deep neural networks to mine and analyze pediatric electronic health records(EHRs) for clinical decision support. Methods: First, we preprocess the unstructured EHRs data in Chinese and transfer them into sentence vectors by natural language processing technologies. Second, we construct the bidirectional recurrent neural networks(BiRNN) model to catch the correlations between patients’ clinical symptoms as well as their interaction with multiple environmental factors. Finally, we train and evaluate our model using a real-world dataset containing 149 817 EHRs. Conclusion: Experimental results show that the proposed method outperforms four baseline methods for clinical decision support accuracy.