Use the Lagrange Remainder Theorem to estimate the error of your approximation.