Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If I have a few values for f(x), i.e. {(0,1), (2, 3), (5, 6)}, is there a way to calculate the derivative at, say f(6), without interpolation?

share|cite|improve this question
up vote 4 down vote accepted

No. You cannot. Given any finite set of coordinates, there is a continuous, nowhere-differentiable function with those points on its graph. Moreover, we can construct everywhere infinitely-differentiable functions with those points on its graph to give any derivative value we like at a given finite set of values. (Both of these claims are readily shown using bump functions, and the former claim also uses the fact that there exist continuous nowhere-differentiable functions.)

We actually need to know the function, or at least know how it is defined on a set having the point we're interested in as a limit point.

share|cite|improve this answer
Right, that was my intuition but I hoped there was something clever out there. Thanks. – mirai Oct 23 '13 at 20:39

In general, no. Even interpolating the data doesn't give you any guarantees that you're capturing the behavior of the function.

If I gave you four nails, could you identify the house that it builds?

Fundamentally, this is because a derivative is defined as a limit... without something to tell you how the limit behaves, you cannot estimate the derivative.

That said, there are things you can do if you want to approximate a derivative if you know the properties of the data. For example, most GPS systems record your driving speed to within 1/2 of a mile per hour (or kilometer per hour). Obviously, your speed is continuous, so if you have a sufficiently large set of GPS speed data, you could approximate the acceleration your car undertakes at any point by attempting to smooth your quantized data.

share|cite|improve this answer
Good answer, thanks. – mirai Oct 23 '13 at 20:40

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.