【作者】 顾琳；

【导师】 夏幼明；

【作者基本信息】 云南师范大学 ， 计算机软件与理论， 2008， 硕士

【摘要】 医学专家系统是人工智能技术应用的一个重要方向,医学诊断领域是医学专家系统的一个核心领域,因为医学的关键在于诊断技术。医学诊断辅助专家系统是医学专家系统在医学诊断领域的推广,它运用专家系统的设计原理,拥有大量专家的宝贵理论以及丰富的临床经验,模拟医学专家诊断疾病的思维过程,协助医生解决复杂的医学问题,可视为医生诊断的辅助工具,甚至能够直接为普通疾病患者提供辅助诊断而不一定需要医生的参与。医学诊断辅助专家系统的信息处理是基于知识智能推理的系统,在功能上它是在医学领域内具有专家水平解决问题能力的系统程序。它涉及到知识获取、知识表示、知识的存储、推理控制机制以及智能人机接口的研究,是集人工智能和领域知识于一体的系统,是一个前景十分广阔的应用领域。本文采用电子病案的形式半自动地获取知识。中医专家对患者进行诊治,与此同时助手或者专家填写电子病案,知识获取程序会自动按照定义好的框架表示形式,将电子病案获得的知识存入文本,然后知识工程师对这些病案知识进行再处理,分别形成知识框架和病案框架,最后将两种知识框架分别存入知识库和病案库。同时具体阐述了基于电子病案和模糊方法的患者自述与标准症状匹配问题。本文通过对本体及其构建的研究,针对中医亚健康领域知识,抽象了一种建立领域本体的形式化方法,并对中医诊断领域知识进行了本体形式化描述与设计。在领域本体的驱动下进行基于中医病案的知识获取,采用基于模糊推理的方法对亚健康症状中医辅助诊断知识进行推理。并在第四章4.5节系统地介绍了基于公理的中医脉诊知识分析与推理。最后本文把以上方法应用到了中医亚健康辅助诊断领域,具体介绍了中医亚健康辅助诊断知识(规则)库、症状库、证候库的分析与设计,并实现了中医亚健康辅助诊断专家系统的证候推理过程。作为人工智能一个重要分支的专家系统(Expert System,ES)是在20世纪60年代初期产生并发展起来的一门新兴的应用科学,而且正随着计算机技术的不断发展而日臻完善和成熟。专家系统是一种智能的计算机程序。这种程序使用知识与推理过程,求解那些需要杰出人物的专门知识才能求解的复杂问题,它能像专家一样解决困难的和复杂的实际问题的计算机(软件)系统。知识表示是为描述世界所作的一组约定,是知识的符号化、形式化或模型化,各种不同的知识表示方法,是各种不同的形式化的知识模型。知识表示的研究既要考虑知识的表示与存储,又要考虑知识的使用。用自然语言与计算机进行通信,这是人类长期以来所追求的。自然语言的识别和处理是人工智能研究的最重要的课题之一,也是人工智能研究的关键。如何去获取各种不同的知识,并以一种计算机可以使用和处理的方法表达知识是知识获取的根本问题。拥用知识是专家系统有别于其它计算机软件系统的重要标志,而知识的质量与数量又是决定专家系统性能的关键因素,但如何使专家系统获得高质量的知识,正是知识获取要解决的问题。知识获取的基本任务是为专家系统获取知识,建立起健全、完善、有效的知识库,以满足求解领域问题的需要。案例学习(CBI)是人工智能中的一种学习方法,该方法由一系列大同小异的学习策略组成,依靠过去的经验进行学习和求解问题.新的案例可以通过修改案例库中与当前情况相似的旧案例来获得。基于案例的推理技术尝试将叙述能力、知识整理进行融合,对有关问题的事件或案例的知识进行萃取。本体是概念化的明确的规范说明。本体可以表示不同的事物：术语表和数据词典,叙词表和分类法,框架和数据模型,形式本体和参考等。一个本体其实就是一个用某种本体语言表达的控制词表,该语言以语法规则限定了词表术语表达具体领域内容的方法,该语法形式上规定了本体控制词表的术语如何共同使用。以详细程度和领域依赖度两个维度作为对本体划分的基础,详细程度高的称作参考本体,详细程度低的称为共享本体。依照领域依赖程度,可以细分为顶级、领域、任务和应用本体等4类。另外,根据主题可分为知识表示本体、通用本体、领域本体、术语本体和任务本体；根据形式化程度分为完全非形式化、结构非形式化.半形式化、形式化的本体。在科学研究和日常生活中,人们一直在追求用一确定的数学模型或康托集合概念来解决问题或表征现象。但专家系统的问题求解一般不象数学、物理等学科那样具有严密性和精确性,它处理的信息往往是不确定的、不精确性的、不完全知道的,甚至是模糊的、不完备的。造成这种现象的原因主要有两点：一是推理依据的规则(或知识)不精确、不完善,而且对不同流派来说还是不一致的；二是证据本身的不确定、不完全甚至有干扰。因此,专家系统设计中不精确的推理使用,几乎是难于避免的,有时成为一个涉及到专家系统设计成败的重要问题。其中有代表性的是如下四种方法：确定性理论、主观Bayes方法、证据理论、模糊集理论。不管是哪一种不精确推理模型,尽管它们处理问题的基本思想和方法有很大差异,但本质是相同的,即都有相同的结构形式,即如下三部分：1)知识不确定性的描述；2)证据不确定性的描述；3)不确定性的更新算法。不精确推理的核心思想是在基于规则的专家系统中,为每个公理本身赋予一个不确定性度量,再给出一组算法,在此基础上,就可以通过这组算法,由公理的不确定性求出定理的不确定性。模糊集理论是一种处理模糊现象的一个极好方法。它多应用于预测型的专家系统中,如经济预测、气象预报、战略布署等。它引起不确定性原因是由模糊性所引起的。它采用隶属函数这种效值计算方法来表达不确定性。其核心思想是要确定诸如：可能性、可能性分布、可能性分布函数、条件可能性分布函数,.边缘可能性分布函数等几个度量和它们之问的关系,以及各种模糊命题的转换规则和不精确命题的推理规则等等。更多还原

【Abstract】 Medical Expert System is an important direction of the application of artificial intelligence technology. The field of medical diagnostic is a core area of Medical Expert System, because medical diagnostic technology is the key of medical field. Medical diagnosis assistant expert system is a promotion of medical diagnosis field,and it uses the design principle of expert system and holds a great deal of valuable expert experience and a wealth of clinical experience.This system can simulate the thinking process of experts diagnosed the disease,and assist to solve complex Medical problems for doctors. It can be seen as an assistant tool of doctor diagnosis,and even could direct provide assistant diagnosis to common patients without doctors paticipating.The information processing of Medical associate experts system is based on the system of knowledge intelligent reasoning. The functions of Medical associate experts system have resolving problem capability in medical field.This system comes down to researchs of knowledge acquisition, knowledge representation, knowledge storage reasoning control and human-computer interface,and it is a system of unity artificial intelligence and domain knowledge,and it is a promising prospect in the application areas.In this paper using the form of semi-automatic to acquisition the electronic medical case. Chinese experts diagnosis and treatment for patients, at the same time assistants or experts fill out the electronic medical case. Knowledge acquisition program could automatically storage these information from electronic medical case according with the form framework into text.After that knowledge engineers dispose them in detail, and respectively creat knowledge framework and cases framework.At last,let those two kind of knowledge storage into knowledge base and case base.At the same time, and expound the matching program of patient’s description and standard symptom based on electronic medical case and fuzzy method.Through the research of ontology and ontology’s constructed,this paper abstract establishment a formal method of domian ontology aims at the domain knowledge of TCM(traditional chinese medicine) sub-health,and makes a formal describing and designing on ontology to TCM diagnosis domain knowledge.Then implements knowledge acquisition on TCM cases by the method of ontology,and bases on fuzzy reasoning method to consequence the knowledge.In section 4.5,make an analysis and reasoning of pulse diagnosis based on TCM axiom method.At last using above methods to the field of TCM sub-health associate diagnosis,introduce the analysis and design of knowledge(rules)base,symptom base and syndrome base of TCM sub-health associate diagnosis system and realize the syndrome reasoning process of a TCM sub-health associate diagnosis expert system. As an important branch of artificial intelligence expert system (Expert System, ES) in the early 1960s and have developed the application of an emerging science, and we are with the continuous development of computer technology and improving and maturing. Expert System is an intelligent computer programs. Such a procedure using knowledge and reasoning process, for those who need the expertise of outstanding personalities in order to solve the complex problems, such as expert as it can solve its problems and the practical problems of complex computer (software) system.Knowledge that the world is described by a group of agreement, is the symbol of knowledge, formal or model, a variety of ways that knowledge is a variety of formal knowledge model. Knowledge that the study should not only consider the knowledge that with the storage, but also consider the use of knowledge.Using natural language and computer communications, this is a long time pursued by mankind. The natural language recognition and artificial intelligence research is handling the most important issue is one of the key study of artificial intelligence. How to access a variety of knowledge and to a computer can use the methods and process of knowledge acquisition of knowledge is a fundamental issue. Yong is the expert system with the knowledge of computer software system is different from other important indicators of, and knowledge of the quality and quantity is the decision of experts the key factor in system performance, but how the experts access to high-quality system of knowledge, knowledge acquisition is to solve Problems. The basic task of knowledge acquisition expert system for access to knowledge and establish a sound, sound, effective knowledge base to meet the needs of solving the problem areas.Case studies (CBI) is the artificial intelligence of a learning method, the method of learning from a series of more or less the same strategy of relying on past experience, learning and solving problems. New case may amend the case with the current situation similar to the old case to get. Case-based reasoning technology will try to describe the ability and knowledge to organize the integration of the issues involved incidents or cases of knowledge extraction.Ontology is the conceptualization of a clear specification. The body can express different things:a glossary and data from the Syrian vocabulary and classification, framework and data model, and other forms of body and reference. In fact, a body is a body language with a word of the control table, the language grammar rules to limit the vocabulary to express specific areas in terms of content means that the grammatical form the bulk requirements Vocabulary control how the terminology used. To the level of detail and dependence on the field as the two dimensions of the bulk of the foundation, the higher the level of detail as reference ontology, the low level of detail as sharing body. In accordance with the degree of dependence on the field, can be broken down into the top, the field, tasks and applications such as body four categories. In addition, under the theme of knowledge that can be divided into the body, common body, the area of the body, body and the task of ontology terms, according to the degree into completely non-formal Formal, non-formal structure. Semi-Formal, the formal body.In scientific research and daily life, people have always been used in the pursuit of a set of mathematical models or Cantor collection concept to solve the problem or the characterization of the phenomenon. However, expert system to solve the problems generally do not like mathematics, physics and other disciplines as a tight and accuracy, processing of information is often uncertain, not accuracy, not fully aware, even vague, incomplete. The cause of this phenomenon are two main reasons: First, reasoning based on the rules (or knowledge) imprecise, incomplete, but also to different schools, or inconsistency and the other is in itself evidence of uncertainty, not entirely or even interference. Therefore, experts in system design using imprecise reasoning, is almost difficult to avoid, and sometimes become involved in an expert system to design the success or failure of important issues. Which is representative of the following four methods:the uncertainty theory, subjective Bayes methods, the theory of evidence, fuzzy set theory. No matter what kind of imprecise reasoning model, even though they deal with the basic ideas and methods are very different, but the essence is the same, that is, have the same structure, namely the following three parts:1) a description of the uncertainty of knowledge 2) a description of the uncertainty of evidence,3) the uncertainty of the updated algorithm.Imprecise reasoning is the core idea in the rule-based expert system, for each of Justice itself gives a measure uncertainty, and then presented a set of algorithms, on the basis of this, we can pass this group algorithm, not by justice Uncertainty obtained theorem of uncertainty.Fuzzy set theory is a fuzzy deal with the phenomenon of an excellent method. It used more than the forecast of expert systems, such as economic forecasts, weather forecasts, strategic deployment, and so on. It is caused by the uncertainty caused by the ambiguous. It uses this function under the validity of calculation methods to express uncertainty. Its core idea is to identify such as:the possibility of possibility, the possibility distribution functions, conditions for the possibility distribution function, the edge of possibility distribution function, and several of the measure and the relationship between them, and various fuzzy proposition conversion rules Proposition reasoning and imprecise rules, and so on. Medical Expert System is an important direction of the application of artificial intelligence technology. The field of medical diagnostic is a core area of Medical Expert System, because medical diagnostic technology is the key of medical field. Medical diagnosis assistant expert system is a promotion of medical diagnosis field,and it uses the design principle of expert system and holds a great deal of valuable expert experience and a wealth of clinical experience.This system can simulate the thinking process of experts diagnosed the disease,and assist to solve complex Medical problems for doctors. It can be seen as an assistant tool of doctor diagnosis,and even could direct provide assistant diagnosis to common patients without doctors paticipating.The information processing of Medical associate experts system is based on the system of knowledge intelligent reasoning. The functions of Medical associate experts system have resolving problem capability in medical field.This system comes down to researchs of knowledge acquisition, knowledge representation, knowledge storage reasoning control and human-computer interface,and it is a system of unity artificial intelligence and domain knowledge,and it is a promising prospect in the application areas.In this paper using the form of semi-automatic to acquisition the electronic medical case. Chinese experts diagnosis and treatment for patients, at the same time assistants or experts fill out the electronic medical case. Knowledge acquisition program could automatically storage these information from electronic medical case according with the form framework into text.After that knowledge engineers dispose them in detail, and respectively creat knowledge framework and cases framework.At last,let those two kind of knowledge storage into knowledge base and case base.At the same time, and expound the matching program of patient’s description and standard symptom based on electronic medical case and fuzzy method.Through the research of ontology and ontology’s constructed,this paper abstract establishment a formal method of domian ontology aims at the domain knowledge of TCM(traditional chinese medicine) sub-health,and makes a formal describing and designing on ontology to TCM diagnosis domain knowledge.Then implements knowledge acquisition on TCM cases by the method of ontology,and bases on fuzzy reasoning method to consequence the knowledge.In section 4.5,make an analysis and reasoning of pulse diagnosis based on TCM axiom method.At last using above methods to the field of TCM sub-health associate diagnosis,introduce the analysis and design of knowledge(rules)base,symptom base and syndrome base of TCM sub-health associate diagnosis system and realize the syndrome reasoning process of a TCM sub-health associate diagnosis expert system. As an important branch of artificial intelligence expert system (Expert System, ES) in the early 1960s and have developed the application of an emerging science, and we are with the continuous development of computer technology and improving and maturing. Expert System is an intelligent computer programs. Such a procedure using knowledge and reasoning process, for those who need the expertise of outstanding personalities in order to solve the complex problems, such as expert as it can solve its problems and the practical problems of complex computer (software) system.Knowledge that the world is described by a group of agreement, is the symbol of knowledge, formal or model, a variety of ways that knowledge is a variety of formal knowledge model. Knowledge that the study should not only consider the knowledge that with the storage, but also consider the use of knowledge.Using natural language and computer communications, this is a long time pursued by mankind. The natural language recognition and artificial intelligence research is handling the most important issue is one of the key study of artificial intelligence. How to access a variety of knowledge and to a computer can use the methods and process of knowledge acquisition of knowledge is a fundamental issue. Yong is the expert system with the knowledge of computer software system is different from other important indicators of, and knowledge of the quality and quantity is the decision of experts the key factor in system performance, but how the experts access to high-quality system of knowledge, knowledge acquisition is to solve Problems. The basic task of knowledge acquisition expert system for access to knowledge and establish a sound, sound, effective knowledge base to meet the needs of solving the problem areas.Case studies (CBI) is the artificial intelligence of a learning method, the method of learning from a series of more or less the same strategy of relying on past experience, learning and solving problems. New case may amend the case with the current situation similar to the old case to get. Case-based reasoning technology will try to describe the ability and knowledge to organize the integration of the issues involved incidents or cases of knowledge extraction.Ontology is the conceptualization of a clear specification. The body can express different things:a glossary and data from the Syrian vocabulary and classification, framework and data model, and other forms of body and reference. In fact, a body is a body language with a word of the control table, the language grammar rules to limit the vocabulary to express specific areas in terms of content means that the grammatical form the bulk requirements Vocabulary control how the terminology used. To the level of detail and dependence on the field as the two dimensions of the bulk of the foundation, the higher the level of detail as reference ontology, the low level of detail as sharing body. In accordance with the degree of dependence on the field, can be broken down into the top, the field, tasks and applications such as body four categories. In addition, under the theme of knowledge that can be divided into the body, common body, the area of the body, body and the task of ontology terms, according to the degree into completely non-formal Formal, non-formal structure. Semi-Formal, the formal body.In scientific research and daily life, people have always been used in the pursuit of a set of mathematical models or Cantor collection concept to solve the problem or the characterization of the phenomenon. However, expert system to solve the problems generally do not like mathematics, physics and other disciplines as a tight and accuracy, processing of information is often uncertain, not accuracy, not fully aware, even vague, incomplete. The cause of this phenomenon are two main reasons:First, reasoning based on the rules (or knowledge) imprecise, incomplete, but also to different schools, or inconsistency and the other is in itself evidence of uncertainty, not entirely or even interference. Therefore, experts in system design using imprecise reasoning, is almost difficult to avoid, and sometimes become involved in an expert system to design the success or failure of important issues. Which is representative of the following four methods:the uncertainty theory, subjective Bayes methods, the theory of evidence, fuzzy set theory. No matter what kind of imprecise reasoning model, even though they deal with the basic ideas and methods are very different, but the essence is the same, that is, have the same structure, namely the following three parts:1) a description of the uncertainty of knowledge 2) a description of the uncertainty of evidence,3) the uncertainty of the updated algorithm.Imprecise reasoning is the core idea in the rule-based expert system, for each of Justice itself gives a measure uncertainty, and then presented a set of algorithms, on the basis of this, we can pass this group algorithm, not by justice Uncertainty obtained theorem of uncertainty.Fuzzy set theory is a fuzzy deal with the phenomenon of an excellent method. It used more than the forecast of expert systems, such as economic forecasts, weather forecasts, strategic deployment, and so on. It is caused by the uncertainty caused by the ambiguous. It uses this function under the validity of calculation methods to express uncertainty. Its core idea is to identify such as:the possibility of possibility, the possibility distribution functions, conditions for the possibility distribution function, the edge of possibility distribution function, and several of the measure and the relationship between them, and various fuzzy proposition conversion rules Proposition reasoning and imprecise rules, and so on.更多还原