0
$\begingroup$

enter image description here

I have the following graph, and i'm trying to determine the maximum. When I use conventional methods, I end up with the ordered pair (2.35, 1.28) roughly. The 1.28 coincides nicely enough, but as you can see, the x values for my graph need to be between 3.6 and 5.1, so this is where I'm basically unsure how to proceed: how do I get the x value between 3.6 and 5.1? Data points for this graph provided on request:

(0.57, 3.6) (1.292, 4.1) (1.058, 4.6) (0.278, 5.1)

$\endgroup$
  • $\begingroup$ Could you provide the four data points used ? There is something strange in the results. $\endgroup$ – Claude Leibovici Oct 8 '15 at 5:05
  • $\begingroup$ @ClaudeLeibovici I added them to the OP body! $\endgroup$ – CKM Oct 8 '15 at 12:03
1
$\begingroup$

Make sure you have a XY scatter plot in Excel, and not a Line plot. Your x values are not used in the parabolic fit you have. In a line plot excel uses values of 1,2,3,4 instead of 3.6,4.1,4.6,5.1.

See the difference below:

line

scatter

Also as a side note, do your analytical processing on the data directly and not on the graph. Unfortunately Excel does not have a polynomial estimator worksheet function (as far as I can tell). You can manually estimate it using a least squares method outlined below:

LSQ

Edit 1

I found a great [SO] answer on how to do a polynomial fit in Excel using undocumented features

$\endgroup$
  • $\begingroup$ I'm embarrassed by this, believe me! Thank you. $\endgroup$ – CKM Oct 8 '15 at 15:01
  • $\begingroup$ Guilty myself. I speak of experience, that's how I spotted the error immediately. $\endgroup$ – ja72 Oct 9 '15 at 12:37
1
$\begingroup$

I really do not see what you did. Using your data points for a quadratic fit, I obtained $$A=-1.502\, V^2+12.8454 \,V-26.187$$ to which corresponds $R^2=0.986743$.

Setting the derivative equal to zero, the maximum corresponds to $V=4.2761$ and, for this value, $A=1.27708$.

I suggest you check how you made the regression.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.