Minimisation of absolute error during calculations with numbers in C++ or C

Question

Is given the function f(x) = x + 1/x. The user inputs let's say "n" integer values x1, x2, ... xn and it is required to output the sum of values of the function for the inputed x-es with and absolute error not more than 10^(-9).
Could u give me a clue or any idea how the method of solving such exercises looks like?

Answer 1

https://medium.com/@dr.sunhongyu/c-linear-regression-example-14243d6b9dc2