摘要
本课题提出一种基于不确定性推理的方法用于《伤寒论》方证知识的量化及挖掘,该技术在《伤寒论》方证量化表达方面的成功应用,为中医方证量化研究提供一种新的思路和方法。
1目的
本课题从研究《伤寒论》方剂入手,尝试利用人工智能技术与知识库技术相结合的方法,通过数学建模的方式实现《伤寒论》方证信息的量化表达,使得学习者通过对处方的分析就能够重现仲景辨证论治过程。
2方法
2.1不确定性推理方法
不确定性推理是指那种建立在不确定性知识和证据的基础上的推理。它是从不确定性的初始证据出发,通过运用不确定性的知识,最终推出具有一定程度的不确定性但却是合理或是近乎合理的结论的思维过程。由于中医学具有模糊性、不确定性和不完全性的特点,导致同一病证的解决方案不唯一,辨证推理所需的知识不确定,专家的背景知识不一致等问题,因此就需要采用不确定性推理方法来研究中医学。
2.2知识库技术
知识库是建立在文献调研的基础上,参考并运用了中医基础理论、中药学、中医诊断学及中医内科学知识而建立的,主要内容有药效知识库、效候知识库、效素知识库、素证知识库。知识库逻辑规则的确定依据中医理论,采用矩阵、条件、规则的方法来表达这些知识,在对《伤寒论》方证的分析研究中,又会产生一些新的认识,可以此来丰富或发展知识库。
2.3算法设计及软件研制
药量是药效作用发挥的基础,也是影响方剂功效改变的重要因素。由于药物的性质不同,其用药的范围会有很大的差异,在方剂中的实际用量很难断定两者在用量上的强度,缺乏可比性。以中医理论为指导,设计具体的数学算法首先对药物的用量进行归一化处理,提出了“药量强度”的概念,使得药物和药物之间的各种药效强度分析有了可比性。同时系统在两药效相互作用强度的计算中,引入“贴近度”的概念：即指两种药物在气、味、归经和药效等因素的综合方面贴近程度,依据同气相求,异气相斥的原理,贴近度越大的药物间的协同作用也越大,反之贴近度越小的药物间的拮抗作用就越大,它可以在计算药效关系时以权重体现,使计算结果更为准确。最后,实现了可在windows系统运行的以方剂为研究切入点的中医处方智能分析系统(CPIAS)。将采集到的《伤寒论》方证信息数据,通过解析标引的方法对其进行一致化处理后即可进入中医处方智能分析系统进行量化分析,从而确定方剂组成各药物的贡献度、方剂综合药气的强度、方剂综合药味的强度、方剂综合归经的强度、方剂各种功效的强度、针对当前病证的临床表现筛选出最适应的功效强度、在筛选出最适应的功效强度的基础上分析与方剂相关的证素、预测方剂的适应证、并确定该方剂所针对的临床表现(症状体征)的主次。
3结果
3.1实现了《伤寒论》方证知识量化表达
本研究通过数学建模与不确定性推理相结合的方法,围绕方剂功效相关因子的各种关系(量效关系、药效关系、效候关系、效素关系、素证关系)的量化计算方法,实现了《伤寒论》112个方证知识的定性定量化表达。分析结果不仅基本符合后世医家对《伤寒论》的普遍认识,而且为解读《伤寒论》方证信息提供了十分有意义的数据支持。通过对《伤寒论》112方药物作用强度的量化表达及排序,实现了《伤寒论》君臣佐使排序；通过对《伤寒论》112方药气、药味、归经的量化排序,发现仲景气味组方规律及方剂归经特点；通过对《伤寒论》112方证的辨证知识要点的量化表达,重现仲景辨证思维过程；通过《伤寒论》非衡量器药物剂量及拆方的研究,拓展了CPIAS的运用范围。
3.2中医知识工程建设的探索
通过研究根据中医理论尝试建立了几千条中医知识,并成功应用于中医处方智能分析系统。虽然这些知识还不能满足中医智能研究的实际需求,但根据此方法和原则的建立的知识库是成功的,可以进一步推广与完善。本课题对建立知识库过程中遇到的一些问题进行了探讨。
4结论
在中医药现代化的发展过程中,引入数学方法,用数学语言描述和表达中医学的知识,实现中医理论的形式化和量化,符合现代科学技术发展的总趋势。通过本课题提出的不确定性推理方法对《伤寒论》方证信息的量化分析和知识挖掘,证明其研究方案是可行的,是数学方法和IT技术深入到中医方药学理论层面的重要实践,有望使这一传统的研究领域发生质的飞跃。
This project presents a uncertain inference method which is used for the quantization and data mining of prescriptions and syndromes knowledge in treatise on febrile diseases.With the successful application of uncertain inference method,it will provides a new idea for the quantitative expression about prescriptions and syndromes knowledge of TCM.
1 Objective
This project starts with prescription study, and uses the method of combing artificial intelligence technology and knowledge base. It will realize quantitative expression ofprescriptions and syndromes knowledge in treatise on febrile diseases by mathematical model and make the learners reproduce the process of syndrome differentiation about zhongjing through the analysis of the prescription.
2 Method
2.1 Uncertain Inference
Uncertain inference is based on uncertain knowledge and evidences. Through the use of uncertain knowledge and initial evidences of departure, it has infered the results which is reasonable or close to the thinking process of a reasonable conclusion. Because traditional Chinese medicine is fuzzy, uncertainty and incomplete features, leading to the same disease and syndrome are not the only solution. dialectical reasoning the knowledge uncertainty, background knowledge of experts and other issues is inconsistent.Therefore, Uncertain inference is needed to study traditional Chinese medicine.
2.2 Knowledge Technology
Knowledge database is based on the review of literature concerned and refer the knowledge of basic theory of TCM, chinese pharmacy, chinese diagnostics and chinese internal medicine. It mainly includes the knowledge of medicine effects, such as pesticide effect Knowledge base, pesticide effect with symptoms Knowledge base, pesticide effect with syndrome elements Knowledge base, syndrome elements with syndromes Knowledge base.The logical rule of Knowledge base is based on TCM theory.Through the quantitative expression of prescriptions and syndromes in treatise on febrile diseases,we can find some new understanding which increase Knowledge base.
2.3 Algorithm Design and Software Development
Dosage is the basis of efficacy, but also affect the prescription effect of the important factors. As the drugs are different in nature, the extent of their drug use would be very different from the actual amount in the recipe,and it is difficult to determine both the amount of the intensity, because of the lack of comparability. Chinese medicine theory as a guide to design specific mathematical algorithm first dosage of the drug were normalized, put forward the "dose intensity" concept, made between the various drugs and drug efficacy has been comparable strength analysis. Pharmacodynamic interaction between the two while the system strength calculation, the introduction of the "closeness" concept.Two drugs in the air, taste, meridian and the efficacy of such factors as the integrated area were closed to the level of gas in accordance with the requirements, different repulsion of the principle of gas, close to the greater degree of synergy between the drugs the greater the contrary, the smaller the close degree of drug antagonism between the greater its efficacy in relationship to calculate the weights reflect, to make the results more is accurate. Finally, realize that can be run in the windows system to study the starting point of Chinese Medicine Prescription Intelligent Analysis System (CPIAS). Treatise on the Prescription and Syndromes information data, by analysis with the method of indexing consistency of treatment to enter prescriptions intelligent analysis system for quantitative analysis to determine the composition of the prescription drug contribution, comprehensive prescription Drug gas intensity, integrated Yaowei prescription strength, prescription comprehensive meridian strength, prescription strength of various effects, for the current clinical manifestations of disease and syndrome screening the effectiveness of the most adapted intensity, in screening the most suitable basis for the effectiveness of strength analysis and prescriptions on the evidence related factors predict prescription indications, and to determine the prescription against the clinical manifestations (symptoms and signs) of major and minor.
3 Result
3.1 Carry out the quantitative analysis of prescriptions and syndromes in treatise on febrile diseases
In this study,mathematical modeling and uncertainty through the combination of methods of reasoning, the relevant factors surrounding the prescription effect relationships(dose-effect relationship, relationship efficacy,effect climate relations, effective relationship factors, prime card relationships) of the quantitative method to achieve the Treatise on the 112 Prescription and Syndromes knowledge,qua-litative and quantitative expression.Basically consistent with later results not only physician on the Treatise on the general knowledge, but also for interpretation, Treatise on the Prescription and Syndromes information provides meaningful data to support. Through quantitative expression of 112 prescriptions with interaction strength,we achieve the ranking of Monarch, Minister,Assistant and Guide in treatise on febrile diseases.Through quantitative expression of four natures, five tastes and channel tropism combined with 112 prescriptions,we found the law of composition with four natures and five tastes, and found the features of channel tropism. Through quantitative expression of syndrome differentiation knowledge combined with 112 prescriptions,we found the dialectical thinking of zhong-jing.Through application of CPIAS in no measurement device and prescription dismantlement research, it will expand the application range of CPIAS.
3.2 The exploration of knowledge engineer of TCM
This study attempts to establish thousands of Chinese medicine knowledge, and is successfully applied in the Chinese Medicine Prescription Intelligent Analysis System. Although the knowledge could not meet the demand of study of Chinese intelligence,but the knowledge base fouded by the method and principle was confirmed successfully and it could be promoted and improved. The project probes a number of issues which was found in research.
4 Conclutsion
The breakthrough point of modernization about Traditional Chinese Medicine Theory is quantitative and digital expression of theory. This is the only way wich could maintain its own characteristics and keep healthy development with the times.Through the quantitative expression and knowledge mining with uncertain inference method used in treatise on febrile diseases,it is proved that research programs is feasible,which has solid foundation and potentiality.
引文
[1]金聪,戴上平,郭京蕾,等.人工智能教程[M].清华大学出版社,2007年1月第1版,97页.
[2]石纯一,黄昌宁.人工智能原理[M].北京：清华大学出版社,1993.
[3]李英华,叶天荣,张红霞.计算机非传统推理导论[M].宇航出版社,1992.
[4]邹开其,徐扬.模糊系统与专家系统[M].西南交大出版社,1989.
[5]张宗明.中医认识中的不确定性因素分析[J].医学和哲学.2000,11,(21) 11：45-47.
[6]王海舜.基于确定性因子的中医辨证不确定性推理模型[J].计算机应用,1999,19(4)：8-111.
[7]秦中广,毛宗源,等.粗糙集在中医类风湿证候诊断中的应用[J].中国生物医学工程学报,2001,20(4)：357-363.
[8]王相东,殷鑫.粗糙集理论与证候规范化研究[J].陕西中医学院学报,2005,28(3)：70-711.
[9]朱文锋,朱咏华,黄碧群.采用贝叶斯网络运算进行中医辨证的探讨[J].广州中医药大学学报,2006,23(6)：449-452.
[10]邰东梅,赵明芳.试论中医诊断的数理智能化[J].中国中医基础医学杂志,2005,11(10)：780-781.
[11]李晓宇,张新峰,沈兰荪.中医舌象信息融合方法的研究[J].测控技术,2007,26(5)：26-28.
[12]李海鲲,宗仁鹤.安徽教育学院学报.基于不确定推理模型的中医诊断专家系统的研制[J].2004,22(6)：39-41.
[13]任廷革,刘晓峰,高全泉,等.中医方荆功效定性和定量研究初探[J].中国中医药信息杂志,2007,14(6)：100—102.
[14]刘晓峰,任廷革,高全泉,等.中医处方智能分析系统的研究与实践[J].中国中医药信息杂志,2007,14(10)：97-99.
[15]高全泉,任廷革,刘晓峰,等.中医方剂药物相对药量改进计算模型[J].海南师范大学学报,2008,21(4)：362-366.
[16]刘晓峰,任廷革,高全泉,等.中医方剂功效定性和定量研究再探[J].中国中医药信息杂志,2009,16(1)：92-94.
[17]任廷革,高全泉,刘晓峰,等.中医方剂功效定性和定量研究三探[J].中国中医药信息杂志,2009,16(6)：96-98.
[18]杨迎霞.论《伤寒论》三阴三阳实质[J].辽宁中医药大学学报.2009,11(1)：17-18.
[19]郭任.《伤寒论》六经病变本质探讨[J].山东中医杂志,2005,24(9)：521-523.
[20]祝建伟,王长美.《伤寒论》关于人体的功能模型[J].山东中医药大学学报,1997,21(3)：178.
[21]徐培平,符林春.《伤寒论》六经营卫观.安徽中医学院学报[J].2000,19(6)：7.
[22]孟庆云.从控制论模糊识别探讨《伤寒论》六经含义[J].陕西中医,1980,(5)：1.
[23]黄宗南,刘家骅.从阴阳逻辑初探《伤寒论》三阴三阳实质—伤寒论数学模型设计[J].北京中医学院学报,1984,(4)：16.
[24]杨玉英.从生化座标系看六经实质[J].山东中医学院学报,1990,14(6)：19.
[25]李秋贵.《伤寒杂病论》的辨证方法研究[J].中国医药学报,2001,6(4)：4-9.
[26]柴瑞震.气血辨证是《伤寒论》的辨证核心[J].中国医药学报,2001,16(3)：11-17.
[27]许爱兰,吴承玉.六经辨证研究思路与方法探讨[J].江苏中医药,2002,23(11)：3-5.
[28]刘玉良.论《伤寒论》定量辨证中的定比思想[J].辽宁中医杂志,2009,(36)1：40-42.
[29]欧阳奇.从伤寒辨证探讨以症定量问题[J].新中医,1991,(1)：11.
[30]杨培坤.运用电脑对《伤寒论》辨证论治思想体系的验证.湖北中医杂志,1981,(3)：1.
[31]熊曼琪,张横柳.临证实用伤寒学.北京：中国科学技术出版社,1991：355.
[32]Sun jizheng, Ren tinge, "Research and Development of Traditional Chinese Prescription Intelligent Analysis Support System"[A],Ed ited by:BARRY BARBER, DEXIAN CAO, DULIE QIN, GUSTAV WAGNER, M EDINFO89 Part2[C], New York:Distributors for the United States and Canada:ELSEVIER SCIEENCE PUBLISHING COMPANY INC, 1989:1144-1147.
[33]李爱敏,王立堂,宋丽军.对现代六经辨证中运用中药配伍量的数理统计分析[J].中医研究,1995,8(5)：9.
[34]王正本.运用集合论对《伤寒论》方剂的研究[J].中医药学报,1984,(5)：19.
[35]包巨太,吴范武,等.《伤寒论》方证对应规律的可视化研究[J].中医杂志,2008,49(5)：372-374.
[36]陈明,陆建峰,赵国平.基于双向强关联规则的仲景桂枝汤类方研[J].南京中医药大学学报.2009,25(4)：249-251.
[37]白晶,盛志刚.几组《伤寒论》经方药效精微的定量计算研究[J].中华中医药杂志,2009,24(5)：644-666.
[38]张建恒.伤寒三阳病病理生理基础假说——中医辨证实质的思考[J].辽宁中医药大学学报,2007,9(5)：45-46.
[39]宋小莉,牛欣,司银楚.基于BP神经网络的半夏泻心汤及其类方配伍模型研究[J].中华中医药杂志,2005,20(6)：383-384.
[40]杨学智,郭宙,司银楚,等.基于BP神经网络的半夏泻心汤及其类方性味药组对小鼠小肠运动的影响[J].中华中医药杂志,2009,24(7)：870-874.
[41]郭继鳌.《伤寒论》电子计算机诊疗和教学系统的研究[J].辽宁中医杂志1982,(4),1-3.
[42]孙华,刘欢中,傅卓.计算机《伤寒论》诊疗软件设计思想[J].北京生物医学工程,1994,13(3)：154.
[43]王天方.刘渡舟教授运用经方的专家经验咨询系统[J].生物医学工程学杂志.1991,8(4)：334.
[44]周庚生,王自勇,王海舜,等.微机《伤寒论》研究文献期刊检索系统的研制和创建[C].首届亚洲仲景学说学术会议论文集,1993：87.
[45]王占彬.《伤寒杂病论》现代临床应用信息系统特色浅论[J].辽宁中医杂志,1999,26(10)：446.
[46]高德,孙华,伍明德,等.《伤寒论》辨证论治系统的研究[C].首届亚洲仲景学说学术会议论文集,1993：21.
[47]刁军成,蒋小敏,等.《伤寒论》病案分析及结果显示计算机教学的实践与总结[J].1998,10(2)：2-3.
[48]王树鹏,艾华,等.《伤寒论》网络教学平台的设计思路与方法[J].辽宁中医药大学学报,2008,10(4)：189-190.
[49]张维波.数学在中医现代化中的意义[J].中国中医基础医学杂志,1998,4(6)：59-62.
[50]陈进.中医的数学建模[J].数理医药学杂志,2002,15(6)：489-491.
[51]中国大百科全书总编委会,黑箱,中国大百科全书(简明版)修订版,中国大百科全书出版社,2004年5月第3版.
[1]杨文潮.论成方中药物作用方向的影响因素[J].陕西中医,2005,26(11)：1227.
[2]朱文锋,晏峻峰.证素辨证新体系的内容及科学意义[J].医学与哲学,2005,(1)：69-70.
[3]朱文锋.创立以证素为核心的辨证新体系[J].湖南中医学院学报,2004,(6)：38-39.
[1]Sun jizheng, Ren tinge, "Research and Development of Traditional Chinese Prescription Intelligent Analysis Support System"[A],Edited by:BARRY BARBER, DEXIAN CAO, DULIE QIN, GUSTAV WAGNER, MEDINFO89 Part2[C], New York:Distributors for the United States and Canada:ELSEVIER SCIEENCE PUBLISHING COMPANY INC, 1989:1144-1147.
[2]高全泉,任廷革,A Measuring Method of Effect Intensity Based on Classified herbs regarding Intelligent Mining of Prescription Efficacies,2009年IEEE计算智能与软件工程国际会议(CISE2009),武汉,2009年12月
[1]刘渡舟.伤寒论校注[M].人民卫生出版社,1991.
[2]李培生,刘渡舟.伤寒论讲义[M].上海科学技术出版社,1985.
[3]高全泉,任廷革,刘晓峰,等.中医方剂药物相对药量改进计算模型[J].海南师范大学学报,2008,21(4)：362-366.
[4]许济群.方剂学[M].上海科学技术出版社,1985.
[5]辛智科.《伤寒论》方证治法分类研究[J].陕西中医,2006,27(5)：620-623.
[6]刘渡舟.新编伤寒论类方[M].山西人民出版社,1984,8(第一版).
[7]黄煌.《伤寒论》类方研究的学术意义[J].南京中医药大学学报,1996,22(3)：29.
[8]杨金萍.《伤寒论》方分类法纵谈[J].中医文献杂志,1999,(4)：3-4.
[9]杨金萍,罗良.《伤寒论》方按八纲分类法之滥觞[J].国医论坛,2005,20(2)：5-6.
[10]杨金萍.《伤寒论》方按病因病证分类法津溯[J].国医论坛,2004,19(7)：1-2.
[11]柯雪帆,赵章忠,张玉萍等.《伤寒论》和《金匮要略》中的药物剂量问题[J].上海中医药杂志,1983,(12)：36-38.
[12]柯雪帆,等.现代中医药应用与研究大系·伤寒与金匮.上海中医药大学出版社,1995,142-145.
[13]梁敬文.《伤寒论》方药剂量古今折算考[J].南京中医药大学学报,1999,15(2)：104-106.
[14]畅达,郭广义.《伤寒论》药物中非衡器计量的初探[J].中成药研究,1985,(8)：44.
[15]孙燕.基于机器学习技术的《伤寒论》方证分析方法研究[D].北京中医药大
学,2007,87-88.
[16]王辉武.经方药物剂量的临床价值[J].中医杂志2002,43(10)：784-785.
[17]李宇航,王庆国,牛欣,等.半夏泻心汤配伍意义的拆方研究一对大鼠离体胃底肌条收缩活动的影响[J].北京中医药大学学报,2000,23(6)：27-29.
[18]刘渡舟.伤寒论诠解.天津：天津科学技术出版社,1983,96.