Towards enhanced hierarchical attention networks in ICD-9 tagging of clinical notes

摘要: Text is an important element in document classification many natural language applications. Natural processing (NLP) today's computational advancement that provides significant modern uses of text documents such as efficient information retrieval. In this paper, we describe the theoretical framework predicting ICD-9 codes through tagging clinical notes using our improved deep learning called EnHANs. This proposed model improvement covers combination word and topic embedding, well adding character-level representation a hierarchical attention neural networks. paper also present use sigmoid activation function last layer enhanced network order to arrive with multi-label, multi-class prediction codes.