Let $f \colon \mathbb{R} \to \mathbb{C}$ be a Schwartz function, for safety.
Is there a bound for $\int_\mathbb{R} \lvert f(x)\ \mathrm dx$ if we know the value of \begin{equation} \int\limits_{\mathbb{R}}\frac{f(x)}{(1+x^2)^n} \ \mathrm dx \end{equation} for some $n \in \mathbb{N}$?
This looks similar to the negative Sobolev norm of index $-n$, but I cannot find a relevant reference for the estimate of $\lVert f \rVert_1$ in terms of the above quantity.
Could anyone help me?
No, this is impossible: for $f\equiv 1$ the integral $\int_{\mathbb R} \frac{|f|}{(1+x^2)^n}$ is finite, while $f$ is not integrable on $\mathbb R$.
Now take $f_k\in C_c^\infty(\mathbb R)$, such that $|f_k|\le 1$ on $\mathbb R$ and $f_k = 1$ on $(-k,k)$. Then $\int |f_k|\to \infty$, while $( \int_{\mathbb R} \frac{|f_k|}{(1+x^2)^n})$ is a bounded sequence.
That is, there is no $C>0$ such that $$ \int_{\mathbb R} |f| \le C \int_{\mathbb R} \frac{|f|}{(1+x^2)^n} $$ for all $f\in C_c^\infty(\mathbb R)$.