If some discrete data points are on a coordinates table, how can I use a equation to fit these points and get some parameters from the equation?
Hi Junhui Li
Matlab can be use. What you should do is minimize residuals between predicted and reals values. You can use optimization app or a script
Non linear regression models are difficult to adjust, specially if they have more than one independent variable.
There some software that allows to do it. But in most cases adjustment are not as good as you can expect.
One of the basics about curve fitting is residuals.
Residuals (yreal - ypredicted)^2 allows to know how good is the fitting made. For example R is based on residuals.
One of the most used techniques to adjust multivariable models is to apply optimization to residuals.
When residuals are minimize difference between real and predicted trends are minimal. This way can be obtain the constants from the non linear model that provides the best possible fit.
- Badges
- Science topic