I have no mathematics background and i appreciate if you provide me with resources to understand which function i should use to best fitting of my data points. Actually, i am using originlab software and i use randomely the gaussian or cube or exponential functions and all of them give good fitting but frankly get to conclusion for my results coz i don't understand these functions.
As far as I know, we can use a particular model for fitting if we can fairly assume that the distribution of data points follows a that specific model. For an example if your data is more like a Gaussian then it will be better to fit a Gaussian. Since in most of the cases the nature of data is unknown, the fitting process is not straightforward. Practically it can be a trial and error experiment. Of course we can calculate the error (fitting deviation) for different models as well as for different parameters (i.e. degrees, mean, variance etc.) in a particular model and finally select a better model (i.e. the model with lowest error and neglect most of the outliers or which does not overfits) which can fairly explain the data. The tool you are using for the task probably tells you the error of each fitted model.
- Badges
- Science topic
- Similar topics
- Mathematics