《应用多元统计分析》PPT课件

《《应用多元统计分析》PPT课件》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、1.        largesampleproblem大样本问题2.        least-squareestimation最小二乘估计3.        levelofsignificance显著性水平[sig'nifikәns]4.        likelihoodfunction似然函数['laiklihud]5.        likelihoodratio似然比['reiʃәu]6.        likelihoodratiotest似然比检验7.        linearrelation线性关系['liniә]8. 

2、       lineartrend线性预测值9.        loading载荷['lәudiŋ]10.logarithms对数['lɔ:gəriðəm]11.lowerlimit下限Newwords12.    MahalDistance马氏距离13.    matrix矩阵['meitriks]14.    maximum最大值['mæksimәn]15.    mean均值16.    meandifference均值差值17.    meansquare均方18.    meansumofsquare均方和19.    meas

3、ure度量['meʒә]20.    median中位数21.    midpoint中值['mid,pɔint]22.    negativecorrelation负相关['negәtiv]23.    nominalvariable名义变量nominal24.    nonlinearcorrelation非线性相关[nɔn'liniә]25.    nonlinearregression非线性回归26.nonparametricstatistics非参数统计['nɔnpærә'metrik]27.nonparametrictest非参

4、数检验28.   normaldistribution正态分布['nɔ:ml]Wehaveseeninthepreviouschaptershowverysimplegraphicaldevicescanhelpinunderstandingthestructureanddependencyofdata.Thegraphicaltoolswerebasedoneitherunivariate(bivariate)datarepresentationsoron“slick”transformationsofmultivariateinform

5、ationperceivablebythehumaneye.Mostofthetoolsareextremelyusefulinamodellingstep,butunfortunately,donotgivethefullpictureofthedataset.3MovingtoHigherDimensionsUnivariate[,ju:ni'vєәriәt]Adj.单变量的Onereasonforthisisthatthegraphicaltoolspresentedcaptureonlycertaindimensionsofthed

6、ataanddonotnecessarilyconcentrateonthosedimensionsorsubpartsofthedataunderanalysisthatcarrythemaximumstructuralinformation.InPartIIIofthisbook,powerfultoolsforreducingthedimensionofadatasetwillbepresented.Inthischapter,asastartingpoint,simpleandbasictoolsareusedtodescribed

7、ependency.Theyareconstructedfromelementaryfactsofprobabilitytheoryandintroductorystatistics(forexample,thecovarianceandcorrelationbetweentwovariables).3MovingtoHigherDimensions→Thecovarianceisameasureofdependence.→Covariancemeasuresonlylineardependence.→Covarianceisscalede

8、pendent.→Therearenonlineardependenciesthathavezerocovariance.→Zerocovariancedoesnotimplyi