电子学报 ›› 2022, Vol. 50 ›› Issue (4): 943-953.DOI: 10.12263/DZXB.20210968

• 机器学习交叉融合创新 • 上一篇    下一篇

一种自适应记忆神经网络多跳读取与覆盖度机制结合的药物推荐模型

王延达1, 陈炜通2, 皮德常1, 岳琳2   

  1. 1.南京航空航天大学计算机科学与技术学院,江苏 南京 211106
    2.昆士兰大学信息技术与电子工程学院,澳大利亚昆士兰布里斯班 4072
  • 收稿日期:2021-07-26 修回日期:2022-03-15 出版日期:2022-04-25 发布日期:2022-04-25
  • 作者简介:王延达 男,1991年出生,山东枣庄人.南京航空航天大学计算机科学与技术学院博士研究生.主要研究方向为数据挖掘及医疗数据分析.E-mail: yandawang@nuaa.edu.cn
    陈炜通 男,1985年出生,广东深圳人.昆士兰大学信息技术与电子工程学院讲师.主要研究方向为数据挖掘、机器学习、医疗数据分析.E-mail: w.chen9@uq.edu.au
    岳 琳 女,1985年出生,吉林长春人.昆士兰大学信息技术与电子工程学院助理讲师.主要研究方向为数据挖掘、机器学习、医疗数据分析.E-mail: l.yue@uq.edu.au第一联系人:
  • 基金资助:
    国家科技创新2030“新一代人工智能”重大项目(2021ZD0113103)

Adaptive Multi-Hop Reading on Memory Neural Network with Selective Coverage Mechanism for Medication Recommendation

WANG Yan-da1, CHEN Wei-tong2, PI De-chang1, YUE Lin2   

  1. 1.College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing,Jiangsu 211106,China
    2.School of Information Technology and Electrical Engineering,University of Queensland,Brisbane,Queensland 4072,Australia
  • Received:2021-07-26 Revised:2022-03-15 Online:2022-04-25 Published:2022-04-25

摘要:

药物推荐的目标是依据病人的电子医疗记录生成药物处方,为医生提供临床决策支持.提取电子医疗记录中蕴含的时序模式以及上下文信息,是成功推荐药物的关键.以往研究忽略了病人之间医疗记录数据量存在差异,无法根据不同病人自身情况,调整数据读取过程中的关注重点以及数据读取迭代次数.针对上述问题,本文提出一种选择性覆盖度机制与自适应记忆神经网络读取结合的药物推荐模型.模型使用记忆神经网络存储病人健康状况对应的时序模式编码结果,利用覆盖度机制进行迭代读取过程中的数据过滤与注意力权重调整.同时模型依据病人自身情况,自适应决定记忆神经网络读取次数.基于真实临床数据的实验结果显示,本模型能够自适应地提取电子医疗记录中的重要数据,构建有效的病人健康状况表示向量,进而完成药物推荐.

关键词: 药物推荐, 记忆神经网络, 注意力机制, 覆盖度机制, 自适应多跳读取

Abstract:

Medication recommendation aims to make effective prescriptions based on electronic healthcare records (EHRs) of patients, and assists caregivers in clinical decision making. Obtaining temporal patterns of patient conditions as well as contextual information contained in EHRs are the key issues for the success of recommendation. Existing methods do not take the difference in the amount of medical records of different patients into account, and fails to change the focus or number of iterations during information extraction according to personalized patient conditions. To address these problems, the medication recommendation model adaptive multi-hop reading with selective coverage mechanism (AMHSC) is proposed. The model stores encoded temporal patterns with memory neural networks (MemNN), and applies the selective coverage mechanism to balance attention weights over selected information during the attentive multi-hop reading on MemNN. Meanwhile, AMHSC adaptively determines the number of reading hops on MemNN according to personalized patient conditions. Experiments on real-world clinical dataset demonstrate that AMHSC successfully derives important information from EHRs to build informative patient representations for medication recommendation.

Key words: medication recommendation, memory neural network, attention mechanism, coverage mechanism, adaptive multi-hop reading

中图分类号: