this is for holding javascript data
Konstantinos Makantasis edited untitled.tex
over 8 years ago
Commit id: 9559aedfb6531bf9115c5f277e99b6503d09ec44
deletions | additions
diff --git a/untitled.tex b/untitled.tex
index 93e2ea5..2bc3775 100644
--- a/untitled.tex
+++ b/untitled.tex
...
\begin{equation}
P\bigg\{\sup_{a\in \Lambda} \bigg (\int Q(z,a)dF(z) - \frac{1}{l}\sum_{i=1}^{l}Q(z_i,a) \bigg ) > \epsilon \bigg \} \xrightarrow[l \rightarrow \infty]{} 0,
\end{equation}
for any $\epsilon > 0$, is called
one-sided uniform
one-sided convergence of means to their mathematical expectations over a given set of functions.
\noindent \textbf{Theorem 1.1} Let there exist the constants $\beta$ and $B$ such that for all functions $Q(z,a)$, $a \in \Lambda$ and for a given distribution $F(z)$ the inequalities
\begin{equation}
...
Then the following two statements are equivalent:
\begin{enumerate}
\item
asd For the given distribution function $F(z)$, the ERM principle is strictly consistent on the set of functions $Q(z,a)$, $a \in \Lambda$.
\item For the given distribution function $f(z)$, the uniform one-sided convergence of the means to their mathematical expectations takes place over the set of functions $Q(z,a)$, $a \in \Lambda$.
\end{enumerate}