MATH 423 Lecture Notes - Lecture 23: Null Hypothesis
Document Summary
Thisis obviously goingto leadproblemif isnt invertible similarly thevarianceof theestimates. If thatmatrixisn"texactlysingular but is close to being nou invertib. ie the variancewill bevery large let 0 . Thereare several equivalent cording any square matrix u to be singular or non invertible. li. 2 at least one eigenvalue of u is 0. 3 one or more x column is equal to a linear combination t the. If wetry to include all of a student"s grades as predictors as well as their overall gpa wewill have theproblem of anearly. 3 u dealwith collinearily by deleting redundant variable hereyoucan either drop all the individual grade or include individualgrades anddrop the average. P suppose we want to test all the lpj jes let sc 1 are o let 5 be be the numberof variables in s forexample for no. 9 3 pis ofpredictors we therefore rejectthe nun hypothesis when. T. n q. in l one more problem about f test.