学科分类
/ 1
3 个结果
  • 简介:Theactivitiesofgeomagneticstormsaregenerallycontrolledbysolaractivities.Thecurrentsolarcycle(SC)24isfoundtobemild;comparedtoSCs19–23,thestormoccurrenceandsizederivedbyaveragingtheoccurrencenumberandDstaroundthesolarmaximumarereducedbyabout50–82%and36–61%,respectively.Weestimateseparately,forSC19to24,therepeatintervalsbetweengeomagneticstormsofspecificDst,basedonfitsofpower-lawandlog-normaldistributionstothestormdataforeachSC.RepeatintervalsbetweensupergeomagneticstormswithDst≤–250nTarefoundtobe0.36–2.95year(s)forSCs19–23,butabout20yearsbasedonthedataforSC24.Wealsoestimatetherepeatintervalsbetweencoronalmassejections(CMEs)ofspecificspeed(VCME)sinceCMEsareknowntobethemaindriversofintensestormsandtherelatedstatisticsmayprovideinformationaboutthepotentialoccurrenceofsupergeomagneticstormsfromthelocationoftheSun.OuranalysisfindsthataCMEwithVCME≥1860km/smayoccuronceper3and5monthsinSC23and24,respectively.BasedonaVCME-Dstrelationship,suchafastCMEmaycauseastormwithDst=–250nTifarrivingattheEarth.BycomparingtheobservedgeomagneticstormstostormsexpectedtobecausedbyCMEs,wederivetheprobabilityofCMEcausedstorms,whichisdependentonVCME.ForaCMEfasterthan1860km/s,theprobabilityofaCMEcausedstormwithDst≤–250nTisabout1/5forSC23or1/25forSC24.AlloftheaboveresultssuggestthatthelikelihoodoftheoccurrenceofsupergeomagneticstormsissignificantlyreducedinamildSC.

  • 标签: solar cycle super GEOMAGNETIC STORM REPEAT
  • 简介:这份报纸与错过反应数据的nonignorable考虑分发函数和quantiles的评价问题。三条途径被开发估计分发功能和quantiles,即,Horvtiz-Thompson-type方法,回归归罪方法和扩充反的概率加权的途径。倾向分数被semiparametric指定指数的倾斜模型。为了在倾向估计倾斜的参数,得分,作者建议一个调整实验可能性的方法处理过去识别的系统。在一些常规条件下面,作者调查asymptotic性质为分发功能和quantiles建议了三个评估者,并且发现这些评估者有一样的asymptotic变化。大折刀方法被采用一致地估计asymptotic变化。模拟研究被进行调查建议方法论的有限样品表演。

  • 标签: 分发 评价 实验 调整 倾斜模型 估计
  • 简介:Inthispaper,wenotonlyconstructtheconfidenceregionforparametersinamixedinteger-valuedautoregressiveprocessusingtheempiricallikelihoodmethod,butalsoestablishtheempiricallog-likelihoodratiostatisticandobtainitslimitingdistribution.Andthen,viasimulationstudieswegivecoverageprobabilitiesfortheparametersofinterest.Theresultsshowthattheempiricallikelihoodmethodperformsverywell.

  • 标签: mixed integer-valued AUTOREGRESSIVE model empirical LIKELIHOOD