How to Calculate Uncertainty in Polynomial Fit Parameters Using Excel?

Click For Summary

Discussion Overview

The discussion centers on calculating uncertainty in polynomial fit parameters using Excel, particularly in the context of non-linear fits. Participants explore methods for obtaining these uncertainties and the relevance of statistical measures like the R² correlation coefficient.

Discussion Character

  • Technical explanation
  • Homework-related

Main Points Raised

  • One participant inquires about methods to calculate uncertainty in parameters from a polynomial fit in Excel, specifically for non-linear fits.
  • Another participant suggests outputting the R² correlation coefficient as a measure of accuracy for the fit, although they express uncertainty about the specific Excel steps to achieve this.
  • A third participant provides guidance on using the "INSERT TRENDLINE" dialog box in Excel, mentioning options for displaying the equation and R² on the chart, along with a caution about decimal precision affecting correlation.
  • The original poster raises a question about fitting data to a specific function, [csc(theta/2)]^4, and whether this topic is covered in numerical methods literature.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best method to calculate uncertainties in non-linear fits, and multiple approaches are discussed without resolution.

Contextual Notes

There are limitations regarding the specific methods available in Excel for non-linear fits and the assumptions underlying the statistical measures discussed.

Who May Find This Useful

Individuals interested in statistical analysis, data fitting, and the use of Excel for mathematical modeling may find this discussion relevant.

JamesJames
Messages
204
Reaction score
0
I performed a fit, using EXCEL, to data and the fit gave me a polynomial. How do I get the uncertainty in the resulting parameters. For example the least squares fit gives y = Ax+B and there are equations to give the uncertainties in A and B. IS there a program that does this for non linear fits?

James
 
Computer science news on Phys.org
You'll want to output the R^2 correlation coefficient. The closer R^2 is to 1, the more accurate the equation is. I know excel can output it, but I'm not sure exactly how since I haven't used excel in a long time.
 
In the "INSERT TRENDLINE" dialog box, there is a tab for options. At the bottom of the dialog box you will see three check boxes. "SET INTERCEPT TO 0," "DISPLAY EQUATION ON CHART" and "DISPLAY R^2 ON CHART." Select the ones you want.

Word of warning, if you check the equation output, you may have to increase the number of decimal places shown in the equation to get better correlation between your regression and the data.
 
How about performing a fit to data that should obey [csc(theta/2)]^4. Would this be out of a numerical methods book or something?

James
 

Similar threads

  • · Replies 28 ·
Replies
28
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
Replies
6
Views
2K
  • · Replies 16 ·
Replies
16
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 30 ·
2
Replies
30
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
8
Views
2K