Rough Sets Model For Multi-granularity Dynamic Hybrid Information Medical Decision-making Problem

Medical management decision-making is the key research direction of "facing people’s life and health" proposed by General Secretary Xi.Scientific and accurate medical and health management program is an important guarantee for life and health.This paper establishes an uncertainty decision-making model and method for the decision-making problem of the whole disease cycle,which focuses on the four stages of clinical diagnosis and treatment whole course cycle: pre-examination and triage,diagnosis decision-making,treatment plan selection decision-making and treatment effect evaluation decision-making,and fully considers the imprecise,hybrid and time-series sequence of modern clinical decisionmaking information,the conflicting and dynamic nature of decision-making attributes that are difficult to portray quantitatively,integrates conflictual large group decision-making theory based on the systematic study of granular computing(GC)frameworks and theories such as multi-granulation rough sets,three-way decision-makings,and formal concept analysis.This paper reviews the characteristics of clinical diagnosis and treatment decisionmaking problems and the differences from traditional decision-making problems,analyzes the shortcomings and difficulties faced by traditional decision-making theories in applying them to clinical diagnosis and treatment decision-making problems,and points out the feasibility and advantages of applying the fusion of multi-granulation rough sets and group decision-making theories to clinical diagnosis and treatment decision-making problems.The TCM clinical decision-making problem and the origins and development of GC framework are reviewed and the current status of domestic and international research is analyzed and compared;the basic concepts and methods used in this paper are introduced.The following aspects of research work has been carried out.1.Constructing a multi-granulation time-series three-way decision-making method for clinical decision-making problems considering conflicting information.The discussion focuses on the clinical dynamic conflict decision-making problem caused by the increase of decision-making information with multi-granulation attribute characteristics and the increase of decision-making subjects with preference difference characteristics in the decision-making process of TCM and Western medicine treatment oriented to actual clinical practice.A decision-making method for integrating TCM and Western medicine based on multi-granulation time-series three-way group conflict analysis is established.Firstly,we describe a portrayal of the conflict information system with multi-granulation preferences,a theoretical model of multi-granulation time-series three-way conflict decision-making is constructed,and its differences and connections with existing related models and its mathematical properties are discussed in detail from the fundamental theory level.Furthermore,an uncertainty decision-making method based on multi-granulation time-series three-way conflict decision-making is given.Finally,the uncertainty decision-making method based on the multi-granulation time-series three-way conflict decision-making is applied to uncertainty decision-making methods based on real-world clinical data for the clinical diagnosis and treatment decision-making of integrated TCM and Western medicine with the characteristics described above.On the one hand,it explains the application steps and basic principles of the uncertain decision-making theoretical model for the context of practical problems of medical decision-making constructed in this paper;On the other hand,the validity and applicability of the theoretical model are verified.2.Building a chronic disease prediction model based on multi-granulation neighborhood three-way clustering.For the clinical chronic disease prediction problems with non-independence of conditional attributes,non-uniqueness of decision-making attributes,and complexity of logical relationships between conditional attributes and decision-making attributes,a chronic disease prediction model based on multi-granulation neighborhood three-way clustering is constructed.Firstly,we consider the correlation between attributes and construct the attribute reduction model based on a multi-granulation neighborhood rough set;we discuss in detail the difference and connection with the traditional neighborhood rough set and its mathematical properties from the basic theory level.Furthermore,we describe the uncertainty decision-making method integrating multigranulation neighborhood rough set,three-way clustering and best-worst method in order to achieve efficient decision-making,minimize the decision-making risk and optimize the decision-making cost.Finally,we apply an uncertainty decision-making method of threeway clustering with multi-granulation neighborhood rough sets and conduct an empirical analysis study based on clinical data of 2683 real-world cases for the TCM chronic disease prediction decision-making problem with the specific characteristics as described above.On the one hand,we illustrate the application steps and basic principles of the uncertain decision-making theoretical model for the context of chronic disease prediction decisionmaking problem constructed in this paper;On the other hand,the validity and applicability of the simulation experimental results of the theoretical model constructed in this paper are empirically verified.3.Introducing a decision-making method for clinical diagnosis of TCM and Western medicine based on multi-granulation rough concept analysis.Taking the TCM and Western medicine clinical diagnosis decision-making problem as the research background,a TCM and Western medicine clinical diagnosis decision-making model based on multigranulation rough concept analysis is developed for the data and knowledge fusion-driven TCM and Western medicine diagnosis decision-making problem in actual clinical practice.Firstly,we describe the portrayal of multi-granulation dominance rough form background with attribute dominance relationship;discuss the method of generating concept lattice in multi-granulation dominance rough form context;construct the theoretical model of multigranulation dominance rough concept analysis fusing data and knowledge,discuss the mathematical properties of the theoretical model in detail from the grounded theory level.Furthermore,we give the uncertainty decision-making based on multi-granulation dominance rough concept analysis method.Finally,we apply an uncertainty decisionmaking method based on multi-granulation dominance rough analysis and conduct an empirical analysis study based on 636 clinical real TCM medical records for the clinical diagnosis decision-making problem of TCM and Western medicine fusion with the characteristics as described above.4.Constructing a multi-granulation time-series conflict decision-making model for clinical treatment scheme optimization.For the problem of clinical diagnosis and treatment scheme optimization decision-making with the incompleteness,dynamics and repeatability of decision-making information generated in the decision-making process,as well as the conflict characteristics between logical relationships,we establish a clinical diagnosis and treatment scheme optimization model based on multi-granulation time-series conflict decision-making.Firstly,we describe the conflict information system portrayal of dynamic directed graphs with incomplete information,and construct the theoretical model of conflict decision-making on generalized incomplete conflict information system;the mathematical properties of the theoretical model are discussed in detail from the basic theoretical level.Furthermore,we describe an uncertainty decision-making method based on multi-granulation time-series conflict decision-making,and use multi-granularity timeseries conflict decision-making method to quantitatively characterize the path dependence of decision-making-makers.Finally,we apply the uncertainty decision-making method based on multi-granulation time-series conflict decision-making and conduct an empirical analysis study based on real clinical data of 2481 of patients for a clinical diagnosis and treatment scheme optimization decision-making problem with the characteristics described above.On the one hand,we illustrate the application steps and basic principles of the uncertain decision-making theoretical model for the context of practical problems of medical decision-making constructed in this chapter;On the other hand,the validity and applicability of the simulation experimental results of the theoretical model constructed in this chapter are empirically verified.5.Introducing a TCM clinical efficacy evaluation model based on time-series dynamic three-way group decision-making.A TCM clinical efficacy evaluation model based on time-series dynamic three-way group decision-making is constructed for the TCM clinical treatment scheme efficacy evaluation problem characterized by the dynamic nature of the decision-making attributes and the dynamic time-series nature of the decision-making process.Firstly,we transform the time-series dynamic group decision-making problem into a three-way group decision-making problem,and the state interval of the decision-making maker is inscribed using the three-way decision-making TAO model with three equal regions,and the state value function is further constructed to evaluate the transfer value of the state region.Furthermore,a three-way decision-making chain is constructed for the state interval transfer decision-making process with time-series dynamic characteristics,and the threeway decision-making chain of group decision-making makers is fused using the decisionmaking cycle,and a time-series dynamic group decision-making information system is established,and the upper and lower approximate definitions of time-series dynamic group decision-making information and the general theoretical definition of the TAO model for group state interval transfer are described to measure the optimal state interval transfer with uncertain strategy and optimal value function,and a new theoretical model of time-series dynamic group decision-making is proposed.Finally,we conducted an empirical analysis using the data of 112 patients with a total of 2376 time points.On the one hand,we describe the decision-making thinking,application steps,basic principles and decision-making process of the uncertain decision-making theoretical model and method constructed in this paper for the context of medical decision-making problems;on the other hand,the validity and applicability of the simulation results of the theoretical model constructed in this paper are experimentally analyzed to further provide theoretical support and auxiliary support for clinical efficacy evaluation decision-making.