When I plot the following data in a xy chart:, the trendline coefficients differ significantly from the results yielded using the linest polynomial equation.
1.692231697901180E+00 1.007320E-01
1.751487512094260E+00 6.969200E-02
1.801485053879390E+00 4.857330E-02
1.860740868072470E+00 2.943180E-02
1.919996682265560E+00 1.891620E-02
1.977400841977050E+00 1.252140E-02
2.032953347206950E+00 8.599360E-03
2.088505852436850E+00 6.218410E-03
2.144058357666750E+00 4.334050E-03
2.205165826341420E+00 3.180580E-03
The 6th order polynomial coefficients (from a6 down to a0) resulting from LINEST are:
3.22981349367398 -28.27012235351990 91.13758396527950 -116.75964325150500 0.00000000000000 126.13519446718700 -78.44804778602400
The 6th order polynomial coefficients (from a6 down to a0) resulting from the trendline are:
-2.465924391528820E+01 2.979221668355000E+02 -1.495759455513340E+03 3.993457659476170E+03 -5.977809562228490E+03 4.754781875397230E+03 -1.569142463507590E+03
Both fits produce good correlation with the original data. If I change the order of the fit to 5th, then the results from linest and the trendline are the same. Also, if I change the data set to the following, the results between the two are the same (for 5th or 6th order).
9.333616979011810E-01 1.007320E-01
1.013258512094260E+00 6.969200E-02
1.080672053879390E+00 4.857330E-02
1.160568868072470E+00 2.943180E-02
1.240465682265560E+00 1.891620E-02
1.317865841977050E+00 1.252140E-02
1.392769347206950E+00 8.599360E-03
1.467672852436850E+00 6.218410E-03
1.542576357666750E+00 4.334050E-03
1.624969826341420E+00 3.180580E-03
Any ideas? This is driving me crazy!
Thanks,
Jay
PS This question was also posted at http://help.lockergnome.com/office/m...124742#3124742 but the answer was determined here first, so don't bother going to the other post.
Bookmarks