基于貼近度的誘導有序加權(quán)平均算子最優(yōu)組合預測模型的構(gòu)建及其性質(zhì).pdf

1、安徽大學碩士學位論文基于貼近度的誘導有序加權(quán)平均算子最優(yōu)組合預測模型的構(gòu)建及其性質(zhì)姓名：袁宏俊申請學位級別：碩士專業(yè)：應用數(shù)學指導教師：陳華友200911AbstractIntheactualforecastproblem，thevariousqualitativeforecastmethodandthequatitativeforecastmethodhavetheirmeritsandtheshortcomingsrespectiv

2、elyThatis，theyalenotmutuallyexclusive，buttheyareconnectedeachotherandtheyhavemutualcomplementationsBecauseeachkindofforecastmethodsusethedifferentdatathedifferentdatasupplythevariousaspectsusefulinformationfromthediffere

3、ntpointIntheforecastingprocess，ifforecastingmakertakesitforgrantedthoughtthatthepredictionerrorisbigforsomesingleforecastmethods，heabandonsthismethodnottouseit，whichpossiblyleadstothelossoftheusefulinformationThecombinat

4、ionforecastingmethodmayobtainbetterforecasteffectscomparedvvimthesingleforecastmethod，andhashigherforecastaccuracythanindividualones，SOithelptodiversifytheforecastriskTherefore，thecombinationforecastingisoneofthetopics、Ⅳ

5、i也researchvalueButthecombinationforecastinghasnotalonghistory，thecombinationforecastingmethodandthevalidtheory’SresearchisnotveryperfeetitisnecessarytofurtherstrengthentheresearchInthisdissertationtheoptimalcombinationfo

6、recastingmodelbasedonapproachdegreeandinducedorderedweightedaveragingoperatorsareconstructedandsomepropertiesarestudiedThemaincontentsareexpressedasfollowing：(1)Theexistingcombinationforecastingalesummarizedinthisdissert

7、ationAnewkindofnonnegativeweighto州amalcombinationforecastingmodelisproposedbasedoncorrelationindexTheapproachdegreeisintroducedtothecombinationforecasting，andthecombinationforecastingmodelisestablishedbasedonarithmeticme

8、anbynon—dimensionalizationofattributeindexesSomepropertiesarestudiedincludingnoninferiorcombinationforecasting，sufficientconditionofsuperiorcombinationforecastingexistence，redundancyforecastmethodexistenceandredundantinf