考虑适应性达成过程的慢性疾病个性化辅助决策方法  

A personalized assistant-decision making method for chronic diseases considering the adaptive reaching process

作　　者：张文博 陈希[1] 张美霞 李贝宁 ZHANG Wenbo;CHEN Xi;ZHANG Meixia;LI Beining(School of Economics&Management,Xidian University,Xi’an,Shaanxi,710071,China;Key Laboratory of Biomedical Information Engineering of Ministry of Education,Biomedical Informatics&Genomics Center,School of Life Science and Technology,Xi'an Jiaotong University,Xi’an,Shaanxi,710049,China;Nursing Department,Xijing Hospital,the Fourth Military Medical University,Xi’an,Shaanxi,710032,China)

机构地区：[1]西安电子科技大学经济与管理学院,陕西西安710071 [2]西安交通大学生命科学与技术学院生物医学信息工程教育部重点实验室,陕西西安710049 [3]空军军医大学第一附属医院(西京医院)护理部,陕西西安710032

出　　处：《系统工程》2023年第4期127-136,共10页Systems Engineering

基　　金：国家自然科学基金面上项目(71974154);陕西省“高层次人才特殊支持计划”青年拔尖人才项目;陕西省自然科学基金资助项目(2020JM-202);西京医院2019年度学科助推计划项目(XJZT19ML60)。

摘　　要：在面向复杂多样的慢性疾病诊断服务过程中,如何根据患者复杂的病情为其提供更加精准、智慧和个性化的辅助诊疗服务,是智能医疗决策的关键问题。基于此,本文考虑针对患者病历开展的案例推理过程中所生成的适应性达成阶段,提出了一种面向慢性疾病的个性化辅助决策方法。在提出的个性化诊疗决策方法中,首先,基于患者复杂多样的诊疗信息,融合相似性和粗糙集理论来确定患者的多种属性权重;其次,提出了自适应规则生成和组合权重偏好学习方法,构建自适应案例推理模型,从海量的历史病历库中为目标病历生成相似的历史病例,从而确定目标病历的患病情况;最后,结合相似性病历的治疗经验和目标病历自身差异化的生理数据,辅助医生为患者提供个性化的治疗方案。本文以糖尿病早期风险预测为例验证该方法的有效性,所提出的自适应规则生成和组合权重偏好学习方法能有效提高案例推理在求解医疗决策问题中的作用,实现复杂诊疗环境下的慢性疾病个性化辅助诊疗。During the process of catering diagnosis services for complex and diverse chronic diseases,how to provide more accurate,intelligent and personalized auxiliary diagnosis services for patients according to their complex conditions is a key issue to promote intelligent medical decision-making.This study proposes a personalized decision-making method for chronic diseases considering adaptive reaching stage in case-based reasoning process for patient medical records.In the proposed personalized diagnosis and treatment decision-making method,firstly,fusing similarity and rough set theory to determine the weight of patients’complex and diverse diagnosis and treatment information;Secondly,the adaptive rule generation and combination weight preference learning methods are proposed,and an adaptive case-based reasoning is constructed to generate similar historical cases for target case from the massive historical case database,thereby determining the disease situation of the target case;Finally,this proposal assists doctors to combine their treatment experiences of similar medical records and the differentiated physiological data of the target patient,providing personalized treatment plans for target patient.In this paper,the validity of the method is verified by an example of early risk prediction of diabetes.The proposed adaptive rule generation and combination weight preference learning method significantly improve the performance of case-based reasoning in solving medical decision problems,realizing personalized diagnosis and treatment of chronic diseases in complex diagnosis and treatment environment.

关 键 词：案例推理 适应性达成过程 组合权重 偏好学习 

分 类 号：C934[经济管理—管理学]