Hello I have some charts which I am trying to fit lines of best fits to.
The general shape of these curves is something representing '4 o' clock' on
a clock.
I am trying to do non-linear regression to get the closest fit to these
curves but the trendline functions on excel are very limited (ie
logarthmic,exponential,polynomial etc) and do not produce a sufficient
trendline.
I was wondering if anyone knows how I could possibly produce the trendline
that I require that fits the data on the curve.
Many thanks for your help.
Warm regards

Nikeel -

If the general shape is a single bulge, and if by "4 o' clock" you mean the
data pattern is an arc from 3 o'clock to 6 o'clock (a southeast bulge), then
the appropriate functions are usually power, exponential, and polynomial of
order 2 (quadratic).

If your data has some other shape, perhaps you could describe it further.

And you might explain what you mean when you say "do not produce a
sufficient trendline."

Also, I recommend that you use an XY (Scatter) chart type.

- Mike
www.mikemiddleton.com

"Nikeel" wrote in message
...
Hello I have some charts which I am trying to fit lines of best fits to.
The general shape of these curves is something representing '4 o' clock'
on
a clock.
I am trying to do non-linear regression to get the closest fit to these
curves but the trendline functions on excel are very limited (ie
logarthmic,exponential,polynomial etc) and do not produce a sufficient
trendline.
I was wondering if anyone knows how I could possibly produce the trendline
that I require that fits the data on the curve.
Many thanks for your help.
Warm regards