I used curve fitting of Excel and I want to know which of them better can explains my data. I know that using R-squared is confronted with some issues in nonlinear curves.
The trouble with what you are doing is that you are comparing curves that have no theoretical basis. The more complex the curve, the better it will fit your data, but the more it will be unreproducible. This is because complex curves will fit the errors in your data as well the underlying function.
I also suggest that, in view of the known statistical errors in Excel, you use a proper stats package. If your data are of any importance, you shouldn't be using Excel.
https://www.washingtonpost.com/news/wonk/wp/2016/08/26/an-alarming-number-of-scientific-papers-contain-excel-errors/
- Badges
- Science topic
- Similar topics
- Philosophy Of Science
- Models