01-09-2012, 09:52 AM
Usually your task is called "fitting", or "approximation". In general for a chosen function you must find such parameters (coefficients) that the sum of the squared errors between the data and the function is minimum. This is called "least squares method". It will be very useful if you know the physical relationship first, transfer it in mathematical equation with coefficients, and then "fit" the equation to your data by changing the coefficients.