一个基于集成情境知识的组合预测方法

《一个基于集成情境知识的组合预测方法》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、311Vol.31,Suppl.1201110SystemsEngineering—Theory&PracticeOct.,2011:1000-6788(2011)S1-0055-11:F272.1;U695.2+2Æ:A12,33,,(1. Æ,100083;2.
,610064;3. ,100190)!#$%,TEI@I&'()%*+,,-.#$/0%1'((CFMIK).

2、2'(/0!34517,8"9#$%,&1:'(%)*+,;<=>,?@A7CDEF1GI-.J/KLM01NOP%92.#QR#3456!17S,TU-CFMIK,AFTER1'(3W8:'1'((ARIMABP-ANN9:XY()%Z&,GI&CFMIK%Z&$;[<41'(.TEI@I;1;/0;#3456!1AcombinedforecastingmethodintegratingcontextualknowledgeHUANGAn-qiang1,XIAOJin2,3,WANGShou-yang3(1.SchoolofEco

3、nomicsandManagement,BeihangUniversity,Beijing100083,China;2.SchoolofBusinessandAdministration,SichuanUniversity,Chengdu610064,China;3.AcademyofMathematicsandSystemsScience,ChineseAcademyofSciences,Beijing100190,China)AbstractAccordingtoQian’smeta-synthesistheoryandTEI@Imethodology,thispaperproposesa

4、combinedforecastingmethodbasedonintegratedcontextualknowledge(CFMIK).Utilizingcontextualknowledgetoguidetheforecastingprocess,thismethodcancovertheinfluenceofthosefactorsthatcannotbeexplicitlyincludedintheforecastingmodel,andthusitcandecreasetheforecasterrorfromstochasticeventstosomeextent.Throughaco

5、ntainerthroughputforecastcase,thispapercomparestheperformanceofCFMIK,AFTER(acombinedforecastingmethod)and3singlemodels(ARIMA,BP-ANN,ExponentialSmoothing).TheresultsshowthattheperformanceofCFMIKisbetterthanthatoftheremainingones.KeywordsTEI@I;combinedforecast;contextualknowledge;containerthroughputfo

6、recast1],Æ,,,. ,, .

.(ARMAARIMAVECMVAR)(Logistic)(ANNSVM).[1],BatesGranger,
.

7、: , .,,!.""!",!
:2011-04-15^_:!#"$"%;!$!#$"&##!#$'$$"&#$`a:!"%,,"&&,%#&!,E-mail: