Dear Kaled,

You can choose curve estimation regression models. To determine which model to use, graph your data. If your variables appear to be linearly related, use a simple linear regression model. When your variables are not linearly related, try transforming your data. When the transformation does not improve things, you may need to use a more elaborate model. Look at a scatter plot of your data. If the plot looks like a mathematical function you recognize, adjust your data to fit that type of model. You also need to determine the quality of the approximation using the coefficient of determination for example. To do this you can use R or any other statistical software.

I hope this helps.

Nassim Mrabti