Problem
f(x)=(e^x-x-1)/(x^2).
When evaluated with |x|<1, the relative error can be large.
a) Explain why this error can be large.
b) Write a Matlab function function y = expmlmx(x) that evaluates f(x) as accurately as possible when |x| < 1. You must use only double precision and must not use any of the Matlab's built-in functions including exp(x).