资源描述：
《基于支持向量机的中长期径流预报分析》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、-------华中科技大学硕士学位论文AbstractRunoffforecastplaysanimportantpartinnotonlywaterrecoursesmanagement,but
alsoeconomicaloperationofhydropowersystem,aswellasthecomprehensiveimpacton
floodcontrol,navigation,waterenvironmentandsustainabledevelopment.Moreoverthe
forecastaccuracyhasadirecteffectonthehy
2、dropowerdistribution.Howeverthe
hydrologicsystemisalargeandcomplicatedsystemandtherunoffelementisinfluenced
byclimate,geographicenvironmentandhumanactivities,etc.Thereafter,itschange
characteristicandrulearecomplicated,random,grayandnon-liner.Thismakesrunoff
forecast,especiallymid-and-long-
3、termrunoffforecasthasbecomecomplicatedand
difficult,thetraditionalmethodsbasedonlineartheoriesaredifficulttomakeagreat
progressinhydrologicforecast.Itisnecessarytodevelopmorenewways.Artificialneuralnetworkhasbeenappliedtohydrologicforecastwidelyduetothe
capabilityofself-learning,self-organi
4、zingandself-adapting,butithasover-learning,local
optimizationandotherissues.However,supportvectormachine(SVM)avoid
over-learningproblemeffectivelybecauseitisbasedonVapnik-Chervonenkis(VC)
dimensiontheoryofstatisticallearningtheoryandstructuralriskminimizationtheory;
SVMadoptquadraticprogram
5、mingandLagrangetheorytogetaglobaloptimization.
Furthermoreitintroduceskernelfunctiontopredigestthesolutionofnon-linerquestion.
Inconclusion,SVMwillbeappliedtorunoffforecast.First,thinkingaboutthedifficultiesofthemid-and-long-termrunoffforecastandin
ordertoimprovetheforecastaccuracy,afterphy
6、sicallyreviewingtheforecastfactors,the
rightoneshavebeenchosenbystatisticallyanalysisandfuzzyoptimummethod.
Second,theforecasthasbeendonebasesontheartificialneuralnetworkandSVM
methodusingtherightfactorsandannuallyrunoffflow.Aftercontrastingandstudyingthe
result,theSVMisfitfortherunoffforec
7、astcomplexityandimprovestherunoffforecast
accuracyinasmallsample.Keywords:Runoffforecast,Forecastfactors,ArtificialNeuralNetwork,SupportVectorMachine-----------II-----------华中科技大学硕士学位论文目录摘要.................................................................
