skewness kurtosis 峰度,偏度 介绍.doc

skewness kurtosis 峰度,偏度 介绍.doc

ID:61786410

大小:130.50 KB

页数:10页

时间:2021-03-20

skewness kurtosis 峰度,偏度 介绍.doc_第1页
skewness kurtosis 峰度,偏度 介绍.doc_第2页
skewness kurtosis 峰度,偏度 介绍.doc_第3页
skewness kurtosis 峰度,偏度 介绍.doc_第4页
skewness kurtosis 峰度,偏度 介绍.doc_第5页
资源描述:

《skewness kurtosis 峰度,偏度 介绍.doc》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库

1、__________________________________________________Skewness,Kurtosis,andtheNormalCurveSkewnessIneverydaylanguage,theterms“skewed”and“askew”areusedtorefertosomethingthatisoutoflineordistortedononeside.Whenreferringtotheshapeoffrequencyorprobabilitydistri

2、butions,“skewness”referstoasymmetryofthedistribution.Adistributionwithanasymmetrictailextendingouttotherightisreferredtoas“positivelyskewed”or“skewedtotheright,”whileadistributionwithanasymmetrictailextendingouttotheleftisreferredtoas“negativelyskewed”

3、or“skewedtotheleft.”Skewnesscanrangefromminusinfinitytopositiveinfinity.KarlPearson(1895)firstsuggestedmeasuringskewnessbystandardizingthedifferencebetweenthemeanandthemode,thatis,.Populationmodesarenotwellestimatedfromsamplemodes,butonecanestimatethed

4、ifferencebetweenthemeanandthemodeasbeingthreetimesthedifferencebetweenthemeanandthemedian(Stuart&Ord,1994),leadingtothefollowingestimateofskewness:.Manystatisticiansusethismeasure_________________________________________________________________________

5、___________________________butwiththe‘3’eliminated,thatis,.Thisstatisticrangesfrom-1to+1.Absolutevaluesabove0.2indicategreatskewness(Hildebrand,1986).Skewnesshasalsobeendefinedwithrespecttothethirdmomentaboutthemean:,whichissimplytheexpectedvalueofthed

6、istributionofcubedzscores.Skewnessmeasuredinthiswayissometimesreferredtoas“Fisher’sskewness.”Whenthedeviationsfromthemeanaregreaterinonedirectionthanintheotherdirection,thisstatisticwilldeviatefromzerointhedirectionofthelargerdeviations.Fromsampledata,

7、Fisher’sskewnessismostoftenestimatedby:.Forlargesamplesizes(n>150),g1maybedistributedapproximatelynormally,withastandarderrorofapproximately.Whileonecouldusethissamplingdistributiontoconstructconfidenceintervalsforortestsofhypothesesaboutg1,thereisrare

8、lyanyvalueindoingso.Themostcommonlyusedmeasuresofskewness(thosediscussedhere)mayproducesurprisingresults,suchasanegativevaluewhentheshapeofthedistributionappearsskewedtotheright.There_____________________________________________________

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。