I am trying to find the Taylor expansion for the function $$f(x) = \frac{1}{1+x^{2}}$$ at $a=0.$ I have looked up the Taylor expansion and concluded that it would be sufficient to show that \begin{eqnarray} f^{(2k)}(0) & = & (-1)^{k}(2k)! \\ f^{(2k+1)}(0) & = & 0 \end{eqnarray}
The problem I am having is that the derivatives of $f$ are horrid - even proving them to be of a particular form seems hopeless. For example, I have tried without success to show that $$ f^{(2k)}(x) = \frac{(2k)!\left( (-1)^{k} + p(x) \right)}{(x^{2}+1)^{2k+1}} $$ where $p$ is a polynomial satisfying $p(0) = p^{\prime}(0) = 0$ and similar for $f^{(2k+1)}.$
I am sure that there is some trick I am missing. The book I am reading hasn't yet covered the binomial expansion for non-integer exponents and so it would seem circular to have to use the binomial expansion for $(1+x^{2})^{-1}.$
How do I derive the Taylor polynomial of order $2n$ at $0$ for $f$?
It is easier to use a geometric series to find the Taylor series. Note that $$\sum_{n=0}^\infty x^n = \frac{1}{1-x}$$
Now substitute $-x^2$ in for $x$ to find:
$$\sum_{n=0}^\infty (-1)^n x^{2n} = \frac{1}{1+x^2}$$
Note that the radius of convergence of the geometric series is 1. That is is works provided $|x|<1$. This also means that the new series is valid provided $|-x^2| < 1$, which means $|x|<1$. Hence it has the same radius of convergence.