Αρχική σελίδα
Τυχαίο
Κοντινά
Σύνδεση
Ρυθμίσεις
Δωρεές
Για τη Βικιπαίδεια
Αποποίηση ευθυνών
Αναζήτηση
Χρήστης
:
Iiirxs/Support vector machine
Γλώσσα
Παρακολούθηση
Επεξεργασία
(
x
→
1
,
y
1
)
,
…
,
(
x
→
n
,
y
n
)
,
{\displaystyle ({\vec {x}}_{1},y_{1}),\ldots ,({\vec {x}}_{n},y_{n}),}
w
→
⋅
x
→
−
b
=
0
,
{\displaystyle {\vec {w}}\cdot {\vec {x}}-b=0,}
w
→
⋅
x
→
i
−
b
≥
1
{\displaystyle {\vec {w}}\cdot {\vec {x}}_{i}-b\geq 1}
, αν
y
i
=
1
{\displaystyle y_{i}=1}
,
w
→
⋅
x
→
i
−
b
≤
−
1
{\displaystyle {\vec {w}}\cdot {\vec {x}}_{i}-b\leq -1}
, αν
y
i
=
−
1
{\displaystyle y_{i}=-1}
.
y
i
(
w
→
⋅
x
→
i
−
b
)
≥
1
,
for all
1
≤
i
≤
n
.
(
1
)
{\displaystyle y_{i}({\vec {w}}\cdot {\vec {x}}_{i}-b)\geq 1,\quad {\text{ for all }}1\leq i\leq n.\qquad \qquad (1)}
max
(
0
,
1
−
y
i
(
w
→
⋅
x
→
i
−
b
)
)
.
{\displaystyle \max \left(0,1-y_{i}({\vec {w}}\cdot {\vec {x}}_{i}-b)\right).}
[
1
n
∑
i
=
1
n
max
(
0
,
1
−
y
i
(
w
→
⋅
x
→
i
−
b
)
)
]
+
λ
‖
w
→
‖
2
,
{\displaystyle \left[{\frac {1}{n}}\sum _{i=1}^{n}\max \left(0,1-y_{i}({\vec {w}}\cdot {\vec {x}}_{i}-b)\right)\right]+\lambda \lVert {\vec {w}}\rVert ^{2},}
[
1
n
∑
i
=
1
n
max
(
0
,
1
−
y
i
(
w
⋅
x
i
−
b
)
)
]
+
λ
‖
w
‖
2
.
(
2
)
{\displaystyle \left[{\frac {1}{n}}\sum _{i=1}^{n}\max \left(0,1-y_{i}(w\cdot x_{i}-b)\right)\right]+\lambda \lVert w\rVert ^{2}.\qquad (2)}
w
→
=
∑
i
=
1
n
c
i
y
i
x
→
i
{\displaystyle {\vec {w}}=\sum _{i=1}^{n}c_{i}y_{i}{\vec {x}}_{i}}
.
y
i
(
w
→
⋅
x
→
i
−
b
)
=
1
⟺
b
=
w
→
⋅
x
→
i
−
y
i
.
{\displaystyle y_{i}({\vec {w}}\cdot {\vec {x}}_{i}-b)=1\iff b={\vec {w}}\cdot {\vec {x}}_{i}-y_{i}.}
w
→
=
∑
i
=
1
n
c
i
y
i
φ
(
x
→
i
)
,
{\displaystyle {\vec {w}}=\sum _{i=1}^{n}c_{i}y_{i}\varphi ({\vec {x}}_{i}),}
b
=
w
→
⋅
φ
(
x
→
i
)
−
y
i
=
[
∑
j
=
1
n
c
j
y
j
φ
(
x
→
j
)
⋅
φ
(
x
→
i
)
]
−
y
i
=
[
∑
j
=
1
n
c
j
y
j
k
(
x
→
j
,
x
→
i
)
]
−
y
i
.
{\displaystyle {\begin{aligned}b={\vec {w}}\cdot \varphi ({\vec {x}}_{i})-y_{i}&=\left[\sum _{j=1}^{n}c_{j}y_{j}\varphi ({\vec {x}}_{j})\cdot \varphi ({\vec {x}}_{i})\right]-y_{i}\\&=\left[\sum _{j=1}^{n}c_{j}y_{j}k({\vec {x}}_{j},{\vec {x}}_{i})\right]-y_{i}.\end{aligned}}}
z
→
↦
sgn
(
w
→
⋅
φ
(
z
→
)
−
b
)
=
sgn
(
[
∑
i
=
1
n
c
i
y
i
k
(
x
→
i
,
z
→
)
]
−
b
)
.
{\displaystyle {\vec {z}}\mapsto \operatorname {sgn}({\vec {w}}\cdot \varphi ({\vec {z}})-b)=\operatorname {sgn} \left(\left[\sum _{i=1}^{n}c_{i}y_{i}k({\vec {x}}_{i},{\vec {z}})\right]-b\right).}
f
(
w
→
,
b
)
=
[
1
n
∑
i
=
1
n
max
(
0
,
1
−
y
i
(
w
→
⋅
x
→
i
−
b
)
)
]
+
λ
‖
w
→
‖
2
.
{\displaystyle f({\vec {w}},b)=\left[{\frac {1}{n}}\sum _{i=1}^{n}\max \left(0,1-y_{i}({\vec {w}}\cdot {\vec {x}}_{i}-b)\right)\right]+\lambda \lVert {\vec {w}}\rVert ^{2}.}
maximize
f
(
c
1
…
c
n
)
=
∑
i
=
1
n
c
i
−
1
2
∑
i
=
1
n
∑
j
=
1
n
y
i
c
i
(
x
i
⋅
x
j
)
y
j
c
j
,
{\displaystyle {\text{maximize}}\,\,f(c_{1}\ldots c_{n})=\sum _{i=1}^{n}c_{i}-{\frac {1}{2}}\sum _{i=1}^{n}\sum _{j=1}^{n}y_{i}c_{i}(x_{i}\cdot x_{j})y_{j}c_{j},}
subject to
∑
i
=
1
n
c
i
y
i
=
0
,
and
0
≤
c
i
≤
1
2
n
λ
for all
i
.
{\displaystyle {\text{subject to }}\sum _{i=1}^{n}c_{i}y_{i}=0,\,{\text{and }}0\leq c_{i}\leq {\frac {1}{2n\lambda }}\;{\text{for all }}i.}
ε
(
f
)
=
E
[
ℓ
(
y
n
+
1
,
f
(
X
n
+
1
)
)
]
.
{\displaystyle \varepsilon (f)=\mathbb {E} \left[\ell (y_{n+1},f(X_{n+1}))\right].}
ε
^
(
f
)
=
1
n
∑
k
=
1
n
ℓ
(
y
k
,
f
(
X
k
)
)
.
{\displaystyle {\hat {\varepsilon }}(f)={\frac {1}{n}}\sum _{k=1}^{n}\ell (y_{k},f(X_{k})).}
[
1
n
∑
i
=
1
n
max
(
0
,
1
−
y
i
(
w
⋅
x
i
−
b
)
)
]
+
λ
‖
w
‖
2
.
{\displaystyle \left[{\frac {1}{n}}\sum _{i=1}^{n}\max \left(0,1-y_{i}(w\cdot x_{i}-b)\right)\right]+\lambda \lVert w\rVert ^{2}.}
[[Κατηγορία:Στατιστική ταξινόμηση]]