Okay so here is my problem. For work, I have a curve that I need to put into
excel and find the equation. So I look at the curve and record data points
from the curve into excel. I will then graph the points and it will produce a
curve just like the one have on paper. My problem comes when I ask excel to
produce equations for the curves. The equations it produces do not match the
curves when I enter a value into the equation. Why does this happen or how
can I fix it?

Make sure you start with an XY chart, not a line chart. Then make sure
the equation has enough digits, so apply a scientific number format with
15 decimal places.

- Jon
-------
Jon Peltier
Peltier Technical Services, Inc.
http://peltiertech.com/


On 4/29/2010 4:23 PM, danklebs wrote:
Okay so here is my problem. For work, I have a curve that I need to put into
excel and find the equation. So I look at the curve and record data points
from the curve into excel. I will then graph the points and it will produce a
curve just like the one have on paper. My problem comes when I ask excel to
produce equations for the curves. The equations it produces do not match the
curves when I enter a value into the equation. Why does this happen or how
can I fix it?

Better lets repeat your procedure. You read out several coordinate pairs on
a curve from some graph. You transferred them into an Excel spreadsheet. You
created a xy-chart. In it, you created a regression line (€œtrend line€�), for
example a quadratic one: y = a0 +a1*x + a2*x2. You read the regression
coefficients a0, a1,€¦ from the text box pertinent to the series of your
points. You tried to write the formula in Excel having the same form as the
regression function. You found out that the results after substituting xs do
not give the expected results. Most probably you are making some basic
mistake (wrong constitution of the equation,+/-, wrong transferred
coefficients etc.). First after eliminating such trivialities you can test
the influence of insufficient accuracy. More transparent method is a direct
statistical treatment of data (LinRegression and similar functions).
--
Petr Bezucha


"Jon Peltier" wrote:

Make sure you start with an XY chart, not a line chart. Then make sure
the equation has enough digits, so apply a scientific number format with
15 decimal places.

- Jon
-------
Jon Peltier
Peltier Technical Services, Inc.
http://peltiertech.com/


On 4/29/2010 4:23 PM, danklebs wrote:
Okay so here is my problem. For work, I have a curve that I need to put into
excel and find the equation. So I look at the curve and record data points
from the curve into excel. I will then graph the points and it will produce a
curve just like the one have on paper. My problem comes when I ask excel to
produce equations for the curves. The equations it produces do not match the
curves when I enter a value into the equation. Why does this happen or how
can I fix it?
.

Rather than copying the trendline parameters, use LINEST to place their
values in worksheet cells. This gives you full precision and the values get
updated should you correct the charting data.
For more see
http://people.stfx.ca/bliengme/ExcelTips/Polynomial.htm
http://people.stfx.ca/bliengme/Excel...&%20LOGEST.htm
best wishes
--
Bernard Liengme
Microsoft Excel MVP
http://people.stfx.ca/bliengme

"danklebs" wrote in message
...
Okay so here is my problem. For work, I have a curve that I need to put
into
excel and find the equation. So I look at the curve and record data points
from the curve into excel. I will then graph the points and it will
produce a
curve just like the one have on paper. My problem comes when I ask excel
to
produce equations for the curves. The equations it produces do not match
the
curves when I enter a value into the equation. Why does this happen or how
can I fix it?