支持向量機技術在華南地區(qū)雷雨天氣預報中的應用.pdf

1、華南理工大學碩士學位論文支持向量機技術在華南地區(qū)雷雨天氣預報中的應用姓名：黃奕銘申請學位級別：碩士專業(yè)：計算機技術指導教師：彭宏鄧文珍20041101華南理工大學碩士學位論文AbstractThunderstormisoneofweatherwhichthreatenbadlytosafetyofaviation.GuangzhouBaiyunairportisoneofthethreebiggestairportsofChinawhi

2、chlocatedinthePearlRiverdelta.ItisthunderstormseasonfromMarchtoSeptembereachyearinthisarea.ThereforeitisveryimportanttotrueforecastweatherofsouthchinaespeciallythunderstormofthePearlRiverdeltaforensuringsafetyofaviationa

3、ndreducingeconomicallylossofairline.SupportVectorMachinehasagooddevelopmentinrecentyears.Therearemanykindsofweatherforecastmethods.HoweverpeoplehavejuststartedtousethemethodsofSupportVectorMachineinweatherforecastinterio

4、rly.ThereforeitisveryimportantthatwestudyweatherforecastwithSupportVectorMachine.Thispaperisstartedwithdatamining.Itintroducesseveralautomaticclassificationmethodsthatwereusedpopularlyatfirst.Lateritresearchesstatistical

5、learningtheoryandprincipleofsupportvectormachinewithdetail.Followingitgoesontoresearchsupportvectormachineinapplicationofthunderstormweatherforecastsouthofchinausingdatapreprocessingmethods.Theresultprovesthatclassificat

6、ionsystemofSupportVectorMachinehashighefficiencyandveracity.Alsotheveracityofresultwillnotfallaccordingtothereductionofquantityoftrainingsample.Itisespeciallysuitableforthemeteorologicalresearchthatisnonlinearandwithshor

7、thistorydatarecorded.Theprincipleofexperienceriskminimizationcouldntmakesureofgoodpredictability.AstosomekernelfunctionwecangetSupportVectorMachinewiththebestperformancethroughadjustingerrorpunishingparameter.Finallythis