100%(2)2 out of 2 people found this document helpful
This preview shows page 1 - 8 out of 28 pages.
Section 3Changwei Xu1/31/2018
Section OutlineIDeletion residuals&Cook’s distanceILinear regression with GLMILogistic Regression R Example
Deletion residuals&Cook’s distanceIWe will review deletion residuals and cook’s distance covered inlast section.IA concrete example will be discussed.
Case Deletion in Linear RegressionUsing the notation from the last section, a subscript (i) means“with the ith case deleted,” for examples:Iˆβ(i)is the estimate ofβcomputed without caseiIX(i)is the(n-1)×pmatrix obtained fromXby deleting theith rowIY(i)is the(n-1)×1 column vector obtained fromYbydeleting theith elementIn particular, thenˆβ(i)=XT(i)X(i)-1XT(i)Y(i)
Deleted ResidualIf we letIyidenote the observed response for theith case, andIˆyj(i)denote the predicted response for thejth case based onthe estimated model with theith case deletedthen theith deleted residual is defined as:di=yi-ˆyi(i)
Studentized ResidualIDeleted residuals depend on the units of measurement just asthe ordinary residuals do. We can solve this problem though bydividing each deleted residual by an estimate of its standarddeviation. That’s where “studentized residuals” come into play.IThe studentized residual is defined asti=yi-ˆyi(i)ˆσ(i)1+xTiXT(i)X(i)-1xiwherexTi(dimension 1×p) is theith row ofXmatrix(dimensionn×p)I*A statistic divided by its estimated standard deviation isusually called astudentized statistic, in honor of W.S.Gosset,who first wrote about the t-distribution using the pseudonymStudent.