Clinical Assistant Diagnosis for Electronic Medical Record Based on Convolutional Neural Network.

摘要: Automatically extracting useful information from electronic medical records along with conducting disease diagnoses is a promising task for both clinical decision support(CDS) and neural language processing(NLP). Most of the existing systems are based on artificially constructed knowledge bases, then auxiliary diagnosis done by rule matching. In this study, we present intelligent approach Convolutional Neural Networks(CNN), which can automatically extract high-level semantic perform automatic without artificial construction rules or bases. We use collected 18,590 copies real-world to train test proposed model. Experimental results show that model achieve 98.67\% accuracy 96.02\% recall, strongly supports using convolutional network learn features conduct assist feasible effective.