摘要
医疗智能诊断推理模型一直是医疗互联网领域的研究重点,有效的诊断推理模型可以帮助医生提高诊断效率,然而疾病的诊断结论需要考虑各种复杂因素,各种类型的疾病应该有自己的诊断模型。本文基于淋巴水肿电子病历提出了一种融合关键字的注意力机制(key-Attention)疾病诊断推理算法,使用电子病历中主诉、家族史、体格检查等内容推理出初步诊断结果。该算法使用的词频-逆文本频率(TF-IDF)算法提取病历关键词,注意力机制选用改进的指针生成神经网络。实验结果表明,该算法能够有效的解决推理模型缺少关键词问题,可以准确地根据电子病历作出疾病诊断推理。
Medical intelligent diagnosis reasoning model has always been the research focus in the field of intelligent medical.An effective diagnosis reasoning model can help doctors improve diagnosis efficiency.However,the diagnosis of diseases needs to consider various complex factors,and various diseases should have their own diagnosis.Based on the electronic medical record of lymphedema,this research proposes a keyword fusion attention mechanism disease diagnosis and reasoning algorithm.In this algorithm,the tf-idf algorithm is used to extract keywords in the medical records,and the attention mechanism uses an improved pointer to generate a neural network.Experimental results show that the algorithm can effectively solve the problem of lack of keywords in the previous reasoning model,and can accurately make disease diagnosis reasoning based on electronic medical records.
作者
王帅帅
徐臻
WANG Shuaishuai;XU Zhen(China Nanhu Academy of Electronics And Information Technology,jiaxing Zhejiang 314000,China)
出处
《智能计算机与应用》
2022年第2期178-181,共4页
Intelligent Computer and Applications
关键词
注意力机制
诊断推理
电子病历
attention mechanism
diagnostic reasoning
electronic medical record
作者简介
王帅帅(1991-),男,硕士,工程师,主要研究方向:nlp、知识图谱。
