Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
using polynomial trendline equations
3 views
Skip to first unread message
john s
Jul 25, 2008, 2:49:32 PM7/25/08
to
I am using trendline feature in charting to develop a third order polynomial
equation. The fit is good and I want to use the displayed equation to
calculate nondata point values.
I cannot get the equation to generate points within the original data set.
Is there a conversion factor I need to use or is there a way to output the
trendline equation differently? The trendline coefficients don't seem to
correspond to the data set even though the graph and r-value look good.
equation. The fit is good and I want to use the displayed equation to
calculate nondata point values.
I cannot get the equation to generate points within the original data set.
Is there a conversion factor I need to use or is there a way to output the
trendline equation differently? The trendline coefficients don't seem to
correspond to the data set even though the graph and r-value look good.
Jon Peltier
Jul 25, 2008, 4:57:05 PM7/25/08
to
1. Use an XY chart, not a line chart.
2. Use as many digits as Excel will provide. Right-click the formula, choose
Format, and on the Number tab, choose a Scientific format with 15 digits
after the decimal.
3. Don't use it for extrapolating. Poly fits go haywire rapidly outside of
the range they were computed over.
2. Use as many digits as Excel will provide. Right-click the formula, choose
Format, and on the Number tab, choose a Scientific format with 15 digits
after the decimal.
3. Don't use it for extrapolating. Poly fits go haywire rapidly outside of
the range they were computed over.
Even better than manually copying the trendline formula and parsing out the
coefficients, at least in Excel 2003) is to calculate the coefficients in
the worksheet using LINEST.
- Jon
-------
Jon Peltier, Microsoft Excel MVP
Tutorials and Custom Solutions
Peltier Technical Services, Inc. - http://PeltierTech.com
_______
"john s" <jo...@discussions.microsoft.com> wrote in message
news:E7238713-43E7-4C31...@microsoft.com...
Tushar Mehta
Jul 26, 2008, 1:57:20 PM7/26/08
to
You may want to check
Trendline Coefficients and Regression Analysis
http://www.tushar-mehta.com/publish_train/data_analysis/16.htm
Trendline Coefficients and Regression Analysis
http://www.tushar-mehta.com/publish_train/data_analysis/16.htm
Regards,
Tushar Mehta
Microsoft MVP Excel 2000-2008
www.tushar-mehta.com
Tutorials and add-ins for Excel, PowerPoint, and other products
0 new messages