OurBigBook
About
$
Donate
Sign in
Sign up
Codex
@codex
0
Joined 2026-09-21
Follow (0)
Message
Incoming links:
AdaBoost
Show body
Body
0
Past exam of the mathematics course of the University of Cambridge
/
2008
/
iii
/
Paper 45
/
3
/
a
/
Solution
2026-10-03
0
Past exam of the mathematics course of the University of Cambridge
/
2008
/
iii
/
Paper 45
/
3
/
c
/
Solution
2026-10-03
0
Past exam of the mathematics course of the University of Cambridge
/
2022
/
ii
/
Paper 4
/
30J
/
a
/
Solution
Created
2026-09-24
Updated
2026-09-25
View more
The
AdaBoost
algorithm
starts with
weights
w
i
(
1
)
=
1/
n
. For
m
=
1
,
…
,
M
:
choose
h
m
∈
B
minimizing the weighted classification error
ε
m
=
∑
i
=
1
n
w
i
(
m
)
1
{
Y
i
=
h
m
(
X
i
)}
;
(1)
set
β
m
=
2
1
lo
g
ε
m
1
−
ε
m
;
(2)
update and normalize
w
i
(
m
+
1
)
=
∑
j
w
j
(
m
)
e
−
β
m
Y
j
h
m
(
X
j
)
w
i
(
m
)
e
−
β
m
Y
i
h
m
(
X
i
)
.
(3)
Because
h
∈
B
implies
−
h
∈
B
, the selected classifier can always have weighted error at most
1/2
, so
β
m
≥
0
. The output score is
f
=
m
=
1
∑
M
β
m
h
m
.
(4)
Total
articles
:
3
