1

In particular to the topic, what does $h_{w,b}(x)=[[\langle w,x\rangle > b]]$ means?

Saar
  • 169
  • Could mean many different things. Does this come from a source? – Eric Towers Dec 16 '18 at 00:51
  • Homework in a course in machine learning. It says "Each classifier is a half space, but the margin doesn’t necessarily go through the origin." (the set in the topic is $\mathcal{H}_d$) – Saar Dec 16 '18 at 00:54
  • Could it be they mean it's just $h_{w,b}(x)=\langle w,x\rangle +b$? – Saar Dec 16 '18 at 00:56
  • Or it might be all the points $x$ such that $\langle w,x\rangle >b$? – Saar Dec 16 '18 at 01:00
  • Ok found the answer in the lecture notes, it's used in the course for indicators, it's 1 when the value in the [[]] is true and 0 otherwise – Saar Dec 16 '18 at 01:19

1 Answers1

0

In the course the lecturer uses it for indicators, which means $[[\langle w,x\rangle >b]]=\begin{cases}1& \langle w,x\rangle >b\\0& otherwise\end{cases}$

Saar
  • 169