3

In

Koenker, Roger, and Ivan Mizera. "Quasi-concave density estimation." The Annals of Statistics (2010): 2998-3027.

There is a notation of adding the symbol $!$ after $\max$ and $\min$

Some examples from the paper:

$$ \prod_{i=1}^n f(X_i) = \max_{f}! \hspace{0.5cm} \mbox{ such that } f \mbox{ is a log-concave density.} $$

$$ \frac{1}{n} \sum_{i=1}^n g(X_i) + \int \psi(g(x)) \mbox{dx} = \min_{g \in \mathcal{C}(X)}! \hspace{0.5cm} \mbox{subject to } g \in \mathcal{K}(X) $$

and one without any constraint:

$$ \frac{1}{n} \sum_{i=1}^n g(X_i) + \frac{1}{|\beta|} \int |g(x)|^\beta \mbox{dx} = \min_{g \in C(X)}! $$

What does that notation mean?

wh0
  • 1,075
  • Could it stand for the objective function? I know German authors like Zeidler also uses that notation – Shamisen Expert Nov 20 '17 at 02:25
  • 3
    It seems to be a shorthand for (in your first example) "is to be maximized with respect to $f$ which is constrained to be log-concave". Sort of like $\text{argmax}$. – kimchi lover Nov 20 '17 at 02:32

0 Answers0