Computes approximation within a given error_limit to the values of a root of an arbitrary function f(x) or to the value of inverse function g(y) = x. Uses Newton-Raphson or Secant methods. Requires one (Newton-Raphson) or two (Secant) estimates of the target value. Raises NoConvergenceError when it does not converge within the set error_limit, on set number of iterations.