You wouldn’t get a better result by curve-fitting then integrating. That would be *less* accurate. When you curve-fit, you’re approximating based on your points. An integral takes points on the curves and uses trapezoids or rectangles to approximate the area; the integral is the limit of those approximations.
Therefore, you’re more accurate to just add up the exact areas of your trapezoids if you have finite discrete data instead of fully continuous data.

The other method stated, to use more points, *would* give a more accurate result.

]]>