欢迎来到天天文库
浏览记录
ID:36631242
大小:2.32 MB
页数:47页
时间:2019-05-13
《探井生产管理决策支持系统知识发现技术的研究》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、中国石油大学(华东)硕士学位论文探井生产管理决策支持系统知识发现技术的研究姓名:陈晓伟申请学位级别:硕士专业:计算机科学与技术指导教师:张琼声20100501KnowledgeDiscoveryTechnologyResearchofExploratoryWellProductionManagementDecisionSupportSystemChenXiaowei(ComputerScience&Technology)DirectedbyProf.ZhangQiongshengAbstractExploratoryWellProductio
2、nManagementDecisionSupportSystemtriedtoapplyartificialintelligencetechniquestoachievecomputerintelligencedecision,whichisusedtoimprovetheveracityandspeedofdecisionofexploratorywellproductionmanagementandreducecostofexploratorywellproduction.Intheprocessofsystemdeveloping,kn
3、owledgediscoverytechniquesareappliedtogainusefulknowledgeandrulefromlargeofexploratorydata,asthesupplementaryofincompleteknowledgeandexperienceofexperts.ThispaperanalyzesdatacharacteristicandfeaturebasedonExploratoryWellProductionManagementDecisionSupportSystem,andchoosesam
4、oreappropriatedatapretreatmentmethod.Thelowaccuratecalculationrate,multi-valueorientationandotherquestionsoccurwhenexistingdecisiontreealgorithmsapplytosystem.Combiningwithroughsettheory,thispaperproposestwodecisiontreealgorithms.OnebasesonattributeweighandtheotherbasesOilr
5、oughclassificationdegree.Twoalgorithmschangeexistingdecisiontreealgorithmsfromdifferentperspectives.Decisiontreealgorithmbasedonattributeweighmakesuseofattributereductiontoweigh,accordingtodifferentimpactondecisionofdifferentattributes.Intheprocessofconstructingdecisiontree
6、,theattributewhichhasthelargestweightisasplittingproperty.DecisiontreealgorithmbasedOffdegreeofroughclassificationconsiderstheimpactOffdecisionresultsofattributeclassificationaccuracyanddependencybetweenattributes.Theexperimentprovesthattwodecisiontreealgorithmseffectivelyi
7、mprovetheclassificationaccuracyanddecisionveracity.Keywords:IntelligentDecisionSupportSystem,KnowledgeDiscovery,Decisiontree,AttributeWeigh,RoughClassificationDegree儿关于学位论文的独创性声明本人郑重声明:所呈交的论文是本人在指导教师指导下独立进行研究工作所取得的成果,论文中有关资料和数据是实事求是的。尽我所知,除文中已经加以标注和致谢外,本论文不包含其他人已经发表或撰写的研究成果
8、,也不包含本人或他人为获得111国石油大学(华东)或其它教育机构的学位或学历证书而使用过的材料。与我一同工作的同志对研究所做的任何贡献均已在论文中作出了明确的说明。
此文档下载收益归作者所有