On strong convergence in \(L^ 1\) for a sequence of Bochner integrable random vectors (Q2366623)

Consider a sequence of random vectors \(X,X_ 1,X_ 2,\dots\) taking values in a Banach space \(E\) with norm \(\|\cdot\|\). Suppose that \(\int(f\circ X)(X_ n-X)dP\) converges to zero in the weak topology defined on \(E\) for every real-valued Borel function \(f\) on \(E\). Then it is shown that \(\{X_ n\}\) converges to \(X\) in mean if and only if \(\{X_ n\}\) converges to \(X\) in law and \(\{\| X_ n\|\}\) is uniformly integrable. Moreover, a particular class \({\mathcal F}\) of real- valued functions is presented such that the convergence in mean of \(\{X_ n\}\) to \(X\) is equivalent to \(\limsup\int F\circ X_ ndP\leq\int F\circ XdP<\infty\) for some \(F\in{\mathcal F}\).