Page 344 - DMTH404_STATISTICS
P. 344

Statistics



                      Notes         Solution.

                                    (a)  The means of X and Y
                                                                                   d
                                         Since the lines of regression pass through the point  X ,Yi , the simultaneous solution of
                                         the given regression equations would give the mean values of X and Y as X =13,Y =17

                                    (b)  The coefficient of determination
                                         We assume that 8X - 10Y + 66 = 0 is the regression of Y on X and 40X - 18Y = 214 is the
                                                                                                              8
                                         regression of X on Y. Thus, the respective regression coefficients b and d are given by
                                                                                                              10
                                             18
                                         and    .
                                             40

                                                                               8   18
                                           The coefficient of determination r  = b.d  =  ´  =  0.36
                                                                       2
                                                                               10  40
                                    (c)  The standard error of the estimate of Y

                                                                  2                               s Y
                                                                                                r
                                         We know that  s Y X  =  s Y  1- r  . To find s  we use the relation  b = ×  .
                                                       .
                                                                            Y
                                                                                                  s
                                                                                                    X
                                                   9       3           b.s X  8   5
                                               2
                                         Also   r =   \   r =   Thus, s =  =   ´   ´ 3 =  4
                                                                    Y
                                                  25       5             r    10  3
                                                         -
                                                s   =  4 1 0.36 =  3.2
                                         Hence,   Y X.
                                    23.6 Summary of Formulae
                                    I.   Regression of Y on X


                                                                         -
                                                                                  å
                                                                                         ( X å
                                         1.   Regression coefficient   b =  å XY nX Y  =  n XY - å  )( Y )
                                                                        2
                                                                                        2
                                                                    å X -  nX 2    n X - å     ) 2
                                                                                           ( X
                                                                                     å
                                                      Cov (X Y,  )  s
                                              Also  b    =      =  r ×  Y
                                                         s 2       s
                                                           X         X
                                         2.   Change of scale and origin
                                                                                  å
                                                   X A          Y B           k  é n uv - å  )( v ) ù
                                                     -
                                                                                        ( u å
                                                                  -
                                              If  u =     and  v =   ,  then  b =  ê             . ú
                                                                                       2
                                                     h            h           h ê  n u - å   ) 2  ú
                                                                                         ( u
                                                                                    å
                                                                                ë                û
                                         3.   Constant term    a =Y - bX
                                         4.   Alternative form of regression equation
                                                                           s Y
                                              Y -  Y =  ( X X-  ) or Y -  Y = ×  ( X X-  )
                                                                         r
                                               C                   C
                                                                           s
                                                                            X
            336                              LOVELY PROFESSIONAL UNIVERSITY
   339   340   341   342   343   344   345   346   347   348   349