It's also called the Lagrange remainder.

Given a Taylor series with n terms, the Lagrange remainder is for some .

What you need to do is to find and pick and so that is maximized. That tells you the maximum error when using the Taylor series with n terms as the approximation.