Polynomial Trendline

[Royi Avital]

Hello.
I found the Trendline function in Charts very useful.
My question is: Are there equivalent functions?

hat I mean, I know about "Trend" and "Linset" functions, yet in Trendline
there are Polynomial Approximations, Exponential Approximations etc... Which
I couldn't find anywhere. Are they available? How can I create the dataset
according to Excel approximation?

Are there anymore Approxiamtions / Interpolations in Excel (Splines, Least
Squares, etc..)?

I'm using Office 2007.

Thanks.

 

[Gary''s Student]

Start here:

http://spreadsheetpage.com/index.php/tip/chart_trendline_formulas/

 

[Gary''s Student]

Start here:

http://spreadsheetpage.com/index.php/tip/chart_trendline_formulas/

 

[Jerry W. Lewis]

As noted by example in the previous link, pretty much all of the chart
trendlines can be fit via LINEST using either array formulas or tranformation
to linearity. A couple of comments are in order, though.

For the transformed fit to be optimal, you must assume equal variance
(across x values) on the transformed scale, not the original scale.
Otherwise either a weighted or nonlinear fit is more appropriate; neither of
which is natively available in Excel, though you could "roll you own" if you
know what you are doing.

Assuming north american regional settings, the polynomial fits shown at the
link assume that the x and y values are given in columns. If they are given
in rows, then the power separater would be a simicolon instead of a comma.

Jerry

 

[Jerry W. Lewis]

As noted by example in the previous link, pretty much all of the chart
trendlines can be fit via LINEST using either array formulas or tranformation
to linearity. A couple of comments are in order, though.

For the transformed fit to be optimal, you must assume equal variance
(across x values) on the transformed scale, not the original scale.
Otherwise either a weighted or nonlinear fit is more appropriate; neither of
which is natively available in Excel, though you could "roll you own" if you
know what you are doing.

Assuming north american regional settings, the polynomial fits shown at the
link assume that the x and y values are given in columns. If they are given
in rows, then the power separater would be a simicolon instead of a comma.

Jerry

 

[Royi Avital]

Are those "Official" formulas by Microsoft?
I see they didn't let the user get the results as a data set.


Thank you very much.

 

[Royi Avital]

Are those "Official" formulas by Microsoft?
I see they didn't let the user get the results as a data set.


Thank you very much.

 

[Royi Avital]

I didn't get your remark regarding the variance of the X values.

Thanks.

 

[Royi Avital]

I didn't get your remark regarding the variance of the X values.

Thanks.

 

[Gary''s Student]

The formulas may not be "official", but they are supplied by a trusted member
of the user community. There is an additional reference you should examine:

http://www.tushar-mehta.com/publish_train/data_analysis/16.htm

 

[Gary''s Student]

The formulas may not be "official", but they are supplied by a trusted member
of the user community. There is an additional reference you should examine:

http://www.tushar-mehta.com/publish_train/data_analysis/16.htm