Thread: Trendlines

When describing a trendline for a polynomial, the curve seems to have good
agreement with the original data, however the displayed equation does not fit
this fitted curve. There is no agreement with the equation and the trendfit
line.

Has anyone else observed this.

The equation displayed on a chart is heavily rounded. Right click on
the equation and format for scientific notation with 14 decimal places.

Another common problem is fitting a trend on a "Line" chart instead of
an "XY (Scatter)" chart. "Line" chart is an extremely misleading name.
It has nothing to do with whether you want to connect the points with
a line or not. Rather, a "Line" chart specifices that x-data (if given
at all) is to be treated as category labels rather than numbers. If you
fit a trend to a "Line" chart, the x-data for that trend will be
1,2,3,... regardless of the user supplied x-data.

You can fit polynomials with LINEST() and TREND() worksheet functions,
but it is easy to create very ill-conditioned polynomial fits, where the
worksheet functions perform poorly compared to the chart trendline fit.

Jerry

Chris wrote:

> When describing a trendline for a polynomial, the curve seems to have good
> agreement with the original data, however the displayed equation does not fit
> this fitted curve. There is no agreement with the equation and the trendfit
> line.
>
> Has anyone else observed this.

Chris -

Probably you are not looking at the equation with sufficient digits.
Select the formula, and use the Increase Decimal button on the
Formatting toolbar, or apply a scientific number format with lots of digits.

- Jon
-------
Jon Peltier, Microsoft Excel MVP
Peltier Technical Services
Tutorials and Custom Solutions
http://PeltierTech.com/
_______

Chris wrote:

> When describing a trendline for a polynomial, the curve seems to have good
> agreement with the original data, however the displayed equation does not fit
> this fitted curve. There is no agreement with the equation and the trendfit
> line.
>
> Has anyone else observed this.
>
>

In news:AADF8899-F8F8-487D-B80B-EDEA8566C952@microsoft.com,
Chris scrive:
> When describing a trendline for a polynomial, the curve seems to have good
> agreement with the original data, however the displayed equation does not fit
> this fitted curve. There is no agreement with the equation and the trendfit
> line.

change the order value in K1
http://www.prodomosua.it/zips/coefficienti.xls
.f