4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院

4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院

ID:18285944

大小:53.50 KB

页数:22页

时间:2018-09-16

4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院_第1页
4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院_第2页
4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院_第3页
4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院_第4页
4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院_第5页
资源描述:

《4 rule-based expert systems - school of computer science and 4基于规则的专家系统-计算机科学学院》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、4Rule-basedExpertSystemsArule-basedexpertsystemisanexpertsystem(seeintro)whichworksasaproductionsysteminwhichrulesencodeexpertknowledge.Mostexpertsystemsarerule-based.Alternativesareframe-based-knowledgeisassociatedwiththeobjectsofinterestandreasoningconsistsofconfirmingexpectations

2、forslotvalues.Suchsystemsoftenincluderulestoo.model-based,wheretheentiresystemmodelstherealworld,andthisdeepknowledgeisusedtoe.g.diagnoseequipmentmalfunctions,bycomparingmodelpredictedoutcomeswithactualobservedoutcomescase-based-previousexamples(cases)ofthetaskanditssolutionarestore

3、d.Tosolveanewproblemtheclosestmatchingcaseisretrieved,anditssolutionoranadaptationofitisproposedasthesolutiontothenewproblem.KBDSSIntro22LucyHedermanTypicalExpertSystemArchitecture(fromLugerandStubblefield)KBDSSIntro22LucyHedermanData-drivenRule-basedExpertSystemsUseForwardChaining:

4、GivenacertainsetoffactsinWM,usetherulestogeneratenewfactsuntilthedesiredgoalisreached.Toforwardchaintheinferenceenginemust:1.Matchtheconditionpatternsofrulesagainstfactsinworkingmemory.2.Ifthereismorethanonerulethatcouldbeused(thatcould"fire"),selectwhichonetoapply(thisiscalledconfl

5、ictresolution)3.Applytherule,maybecausingnewfactstobeaddedtoworkingmemory4.Haltwhensomeuseful(orgoal)conclusionisaddedtoWM(oruntilallpossibleconclusionshavebeendrawn.)KBDSSIntro22LucyHedermanGoal-drivenRule-basedExpertSystemsUseBackwardChaining:Workbackwardsfromahypothesisedgoal,att

6、emptingtoproveitbylinkingthegoaltotheinitialfacts.TobackwardchainfromagoalinWMtheinferenceenginemust:1.Selectruleswithconclusionsmatchingthegoal.2.Replacethegoalbytherule'spremises.Thesebecomesub-goals.3.Workbackwardstillallsub-goalsareknowntobetrue-eithertheyarefacts(inWM)ortheuser

7、providestheinformation.KBDSSIntro22LucyHedermanExampleAproductionsystemIDENTIFIER,whichidentifiesanimals.R1IFtheanimalhashairTHENitisamammalR2IFtheanimalgivesmilkTHENitisamammalR3IFtheanimalhasfeathersTHENitisabirdR4IFtheanimalfliestheanimallayseggsTHENitisabirdR5IFtheanimalisamamma

8、ltheanimaleatsmeatT

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。