An Applicationof SoftComputing to RETS: Rheumatic Evaluation and Treatment System by OrientalRheumaticMedicin
Biomedical Fuzzy BiomedicalFuzzySystems Systems Association Association SIIg BMFsA(2oo6-i2-oi-o2) BiomedicalSoft Cornputing and Human Seiences,VoL.11, No.1, pp.7-13 Systems Assoeiatien Copyright @ 1995 Biomedecal Fuzzy Con tn'buted Articles An Applicationof SoftComputing to RETS: Evaluation and Treatment System by OrientalRheumatic Medicine Thang CAO",Eric W COOPER"", and " Yukinobu HOSHINO"**, Hoang Phuong NGUYEN""""Katsuari GraduateSchoolofScience andEngineering KAMEI"* RitszttneikanCiniversi4･' Seience andEngineering Ritsumeikan(Jhiversity of1nj/ormation andPhotonic EngineeringKochi University qfElectronic S)]siem of7lechnology ""'* IFIformation7leehnoiogyCenter, Mlnistry qf'Health ofVietnam **Cotlege *"Dept 15,200bl treceivedApril Abstraet: in thispaper, wepJ'esent an [uzpiication eomputing intoan erpert system RE7S: Rheumatic ofsoj7 Evatuation and 1>'eatment O}'iental A4edieine Imputs are severities ofobserved symptoms Systemby (Oltd7 on patientsand ouiputs are a cfiagnosis oj'rheumatie states, itsexplanation andherbalprescriptions, First, clfrer usingjuzay iizferences, RE7S diagnoses the most suitable rheumatic state in which thepatient c\lpears to be in.fected and shows explanations Nexc bciyed on prescriptionsfor rheumatism learned by neural networks,77'om skitted OMpltysicians and OM text books, RE7S gives an orientat prescriptionwritten in suitabie herbswith recxsonabte amozints Finally,we clescribe evaluations and restrictions and ofRE7:Sl then describe ourfature work Keywords: expert system decision support system, neura] networks, Introduction Rheuunatism isan arthritis diseasewidespread in unfortunately populationgroups, influencing socioeconomic aspects of Vietnam It accounts for 15% of all soft tissuediseases The most type, joint degeneration, accounts popular rheumatic fbr 109'6 [1].In Victnam, therapeutic ways fbr Vietnamese a]l rheumatism physical rnethods, are anti-infiammatory (OM).Among them, OM isan indispensable because ithas fewer side-effects part than western medicine and gives good treatment results Besides,herbalprescriptions are easy to find and oriental medicine and relatively drugs.The cheap number in comparison with western patientstreated by ofVietnamese isabout 50 % The accurate diagnoses have an important role in disease treatrrients Building a successfu1 expert OM RETS based on knowledge from wiLl he]p moderate evaluation physicians inrheumatic diagnoseswhich tend to be sub.jective It will indirectlyhelp physicians to providethe right system such as OM skiiled treatments to the right patients,thus improving the qua]ityof the health care services as a whole Italso help qualifiedand experienced physicians in OM to maintain and share their profound knowledge with col]eagues and to assist medica] students or young those living and working in physicians, especially rural areas 1-1-1 Noji-Higashi, Kusatsu-city, Shiga-prefecture, 525-8577, Japan Phone and Fax: +81-77-561-2861 fuzzyinference Expert System (ES) In the lastrecent decades,the advent of the has greatly stimulated developmentsof ES which perfbrmsthe roles of a specialist or carry out some works requiring specific expertise There were many domains in which ES has been successfu11y computer applied, chemistry such as and [7] According to OM, disease states [2-4],geology [6], medicine business[81 and rheumatism 32 mainly consists of I2 typiealclinical The number of herbs for rheumatic treatment is63 [11].Based on severities of observed symptoms on patients,doctors diagnose and classify states of rheumatism, then givecorresponding herba] amounts in grams.Fig prescriptionswith reasonable shows the processofdiagnosis and prescriptionof rheumatic treatment by OM doctors Such a process can be suitably assisted with an ES as shown inFig.2 Functional ofFig.2 are as fo1[ows: [9] parts Knowledge Acquisition:Survey symptoms, symptoms prescriptionsand training data Base: Consists of symptoms, inference rules, trainingdataand explanations Fuzzy Inference: Checks rules calculates weights and giyesproper rheumatic state Inputs to the fuzzy inferenceare seyerities of observed symptoms and ruEes, and one output that isthe most serious ofthe theumatic state that the patient has explanations, sample Knowledge Neural Network: Gives reasonable amounts of herbsintreatment prescriptions Inputsto NN are also severities of observed symptoms, and outputs from NII-Electronic Library Service Biomedical BiomedicalFuzzy Fuzzy Systems Systems Association Association Biomedical Soft Computing NN are coefficients Human SciencesVol.11 and in [O, 1] amounts ofherbal No.1.December2006 They and thejr symptoms severities from users and shows resutts written in infectedrheumatic states and appropriate treatment Interface: Obtains prescrlptlons Explanation: rheumatism, ES Helps users and explains understand OM, fromRETS results : Kti POi,i'ikE/,:.,i,//l.: [ "TiONS developersto use linguistic builda friendlyuser interface OM also enable and variables physiciansusually with symptoms typical with those those can herbswith be represented In addition, and `Cthis These expressions weights." in IF-THEN fuzzy rules giye expert-like it easier for doctorsto rules can make ES the understand as these naturally fuzzy explanations, dlagnosing procedures patienthas these severities, so I prescribe explain expressions such So far,based on fuzzy logicmany practical , ･11///////lt,f,･.//g///h･/,･,i･//, Pulseexffniination applications in medicine have been built, including Dlagnose OM [2i and rheumatic disease with western medicine ± -[4] In the first stage of RETS, fuzzy inference is Adiseasestate[sdete[mned cjin/ea]expeFiences states the patienthas used to decidewhich rheumatic "' Then in the second stage, the fuzzy severities of Consdderasu;table neural symptoms are put into a corresponding Herbaludiustrnents l standaTdpFescriptiOn network to get an appropriate prescription ' FINALPRESCRIPTION 3.1.Symptom and Rule Expressions ll:i:L//1:'/1/iYgrrMn" Herbkl.'.1.'. -.gra71 Fig.1 Diagram ofdiagnosis and treatment by OM rheumatic has m==32 clinical Suppose that rheumatism states A rheumatic state l 12 rheumatic is determinedby n clinical symptoms, Sinceall of the clinical symptoms will not appear in one rheumatic state, the value of n may yary depending 16 forstate 1, n I3 on each state, fbr example n symptoms, prescriptionof doctors = =: fbr statc 2, and Let S" so == on the =(S,O,S,", ,S:i)be symptoms on patient a representmg proposltlon si{J be the be ==(s]RJ,s2RJ, s,R,J) IF Sfi'andS2R' THEN of the of rule a set of RJ-(J' 1,2, ,l) the and ,andS.it premise of (1) isIIr rheumaticstate Rule fbrm (1)means as set R.iisgenerally described inthe fo11owing fbrrn: where ES for RETS fuzzy a a symptom in the premise symptoms of observed states Let Fig Structure of S,O･ is where Let H=(Hi,H2 lli) rheumatic set there are in symptoms n Symbol S,Rican be understood the rule flytTuptom1 in premise of ru]e Rj.j is rule number 3.Fuzzy Inference rn Let the two fbllowingfuzzy yalues OM, give herba] physicians usually basing on severities of clinical prescriptions symptoms such as high fever,slightly numbed powerfu1 yellow approaches ' urine to etc resolve the above Si'bedefined: ltsf)E joints, These fuzayexpressions of symptorns make it unsuitable for traditional approaches to buildOM expert systems quantitative Fuzzy sets, known fbr their abilities to deal with vague variables using membership functionsrather than with crisp vaiues, proyedto be one of the most moderately in S,O and when on on [O,1]i truthvalue diagnosing "s,o =1 the patient,/ts,e =O the patient,and of S,O given by doctors S,O means clearly S,C･) does not means O < ps;/ < means appears appear S,O･ appears on the patient with the severity "sf)･ problem NII-Electronic Library Service Biomedical Fuzzy BiomedicalFuzzySystems Thang Cao Systems Association Association Applieationof Seft Cemputing to RETS: Rheumatic Evaluation and et al.:An E[O,1] ltsft, value Hy given by state rheumatic importance : R Si' of doctorsvia skilled SR' symptoms SO symptoms conclusion , RJ match of weight by u,R, )iscalculated (denoted in advance, wherc: survey If for Treatment System by OrientalMedieine observed the of rule RJ as: n l.]p,,1j ltsn, O means = (2) S,iCJ has absolutely no affect on S,eJjs the symptom H,i, and O