Page 325 - Excel for Scientists and Engineers: Numerical Methods
P. 325

302                                        EXCEL: NUMERICAL METHODS


                   After formatting to show a few more decimal places, for example,
                   y = -1.72727E-04x3 - 4.94605E-03x2 - 5.38589E-01~ 3.20986E+01
                                                                   +
               you are now ready to copy the values and paste them into your spreadsheet.

               Importing Trendline Coefficients into a Spreadsheet
               by Using Worksheet Formulas
                   The following are the various Trendline fitting functions that are displayed in
               the Add Trendline dialog box directly into worksheet cells:


                              linear                       y=ax+b
                              logarithmic                  y = a In(x) + b
                              polynomial (e.g., order 3)   y = ax3  bx2 + ex i d
                              power                        y = ax  b
                              exponential                  y = aebx


                   The  linear,  logarithmic  and  polynomial  expressions  are  linear  in  the
               coefficients  and  can  be  handled  by  Excel's  built-in  linear  regression  code.
               Trendline uses linear transformation of the power  and  exponential  functions to
               obtain the coefficients: the exponential expression is transformed to lnb) = bx +
               In(a) and the power expression to Inb) = b In@) + In(a).
                   The  following  formulas  allow  you  to  get  the  coefficients  of  the  various
               Trendline fitting functions directly  into worksheet  cells.  The formulas use the
               results  returned  by  LINEST,  so  there's  really  no  reason  not  to  use  LINEST
               directly.  But  for those  die-hards  who  insist  on  using  Trendline,  here  are the
               relationships (in each formula, replace the arguments y-values  and x-values  with
               the appropriate range references):

                   linear        a   =INDEX(LINEST(y-values,x-values,l  ,O),l)
                                 b  =INDEX( LINEST(y-values,x-values,  1,0),2)


                   logarithmic   a   =INDEX( LI NEST(y-values, LN(x-values), 1 ,O), 1 )
                                 b   =I  N D EX( LI N EST( y-values,  LN(x-values) ,I ,0) ,2)

                   polynomial    a   =INDEX(LINEST(y-values,x-valuesA{l ,2,3},1 ,O),l)
                   (e.g., order 3)  b  =INDEX(LINEST(y-values,x-valuesA{l ,2,3},1,0),2)
                                 c   =I  N D EX( L I N EST( y-val ues, x-valuesA{ 1 ,2,3}, 1 , 0) ,3)
                                 d   =I  N D EX( LI N EST( y-val ues, x-valuesA{ 1 ,2,3}, 1 , O), 4)
   320   321   322   323   324   325   326   327   328   329   330