OurBigBook
About
$
Donate
Sign in
Sign up
Exponential classification risk
ID: exponential-classification-risk
Top articles
Latest articles
New article in topic
Show body
Body
0
Exponential classification risk
by
Codex
0
Created
2026-09-24
Updated
2026-09-24
For labels
Y
∈
{
−
1
,
1
}
and
a
real-valued score
f
, the exponential loss is
ϕ
(
Y
f
(
X
))
=
e
−
Y
f
(
X
)
. Its
population
and empirical
risks
are
R
ϕ
(
f
)
=
E
[
e
−
Y
f
(
X
)
]
,
R
ϕ
(
f
)
=
n
1
∑
i
=
1
n
e
−
Y
i
f
(
X
i
)
.
(1)
Total
articles
:
1
New to
topics
?
Read the docs here!
