Skip to main content
Haiyue's Blog
Blog Home
Techniques
Study
Work
Models
Search
Ctrl
K
Distributions
Haiyue
September 15, 2024
Less than 1 minute
math
s
i
g
m
o
i
d
(
v
a
l
u
e
,
f
a
c
t
o
r
=
100
)
=
(
1
1
+
e
(
−
v
a
l
u
e
/
f
a
c
t
o
r
)
−
0.5
)
∗
2
sigmoid(value, factor=100) = (\LARGE{\frac{1}{1+e^{(-value/factor)}}}-0.5)*2
s
i
g
m
o
i
d
(
v
a
l
u
e
,
f
a
c
t
or
=
100
)
=
(
1
+
e
(
−
v
a
l
u
e
/
f
a
c
t
or
)
1
−
0.5
)
∗
2
W
i
=
1
N
W_i = \frac{1}{N}
W
i
=
N
1
Start X:
End X:
Factor:
The formula
(1/(1+(Math.E**(-x/this.factor)))-0.5)*2
Start X:
End X:
Start Y:
End Y:
Weight:
The formula
/* sigmoid(value){ let factor = 18 return (1/(1+(Math.E**(-value/factor)))-0.5)*2 } */ this.weight*(this.sigmoid(x) + this.sigmoid(y))
Reference
Sigmoid function
Prev
Confirmatory Factor Analysis