Is there a way to compute some unknown parameters in an equation using curve fitting using experimental data?
I am upload a picture which contains the function and three unknown parameters and it says that the author used minimizing sum the sum of squared residuals method to calculate the values of the parameters.
@ Lukas Wesemann
Yes you can use curve fitting to extract the unknown coefficients of an equation.
If you have three parameters then you can choose 3 different points on the curve ,
Then you can build three equations which you can solve to get the unknown parameters. Then you can get the values of y in the other points and see how much is the error.
For the solution of the equation as they are nonlinear you can get out the parameters values which gives MMSE between ymeasured - y estimated from the coefficients.
Best wishes
Yes you can use curve fitting to extract the unknown coefficients of an equation.
If you have three parameters then you can choose 3 different points on the curve ,
Then you can build three equations which you can solve to get the unknown parameters. Then you can get the values of y in the other points and see how much is the error.
For the solution of the equation as they are nonlinear you can get out the parameters values which gives MMSE between ymeasured - y estimated from the coefficients.
Best wishes
SAS has an NLIN method where you can specify the model as shown and it will using least squares compute the coefficients along with their confidence intervals.
Dear Abdelhalim abdelnaby Zekry , thank you for your response. If you could help me further I would like to know as to how use MMSE to get out the parameter values from the non linear equation that I have. Is the procedure same as you mentioned to use three data points and substitute the values of x and y, which in my case would be t and J, the simultaneously solve to find the parameter values?
How would have to use MMSE?
Dear Srini Vasan , Thank you. All right I will look into it.
Dear Eng. Sayed
use curve fitting, usually with least square minimization and choose a third order polynomial. y = ao + a1x +a2x^2 The 3 coefficients will give you the three parameters values, i your equation is algebraic. The algorithm of Polynomial Regression exists everywhere.
Dear Muhammad Hamza El-Saba , all right that I will try. I think that would also be possible with MS Excel. Graph options -> trendline->polynomial (order=3). Any other software that you would recommend?
- Badges
- Science topic
- Similar topics
- Engineering
- Computer Engineering
- Computer