平稳随机过程
P r { X 1 = x 1 , . . . , X n = x n } = P r { X 1 + l = x 1 , . . . , X n + l = x n } Pr\left\{ X_{1}=x_{1},...,X_{n}=x_{n} \right\}\\=Pr\left\{ X_{1+l}=x_{1},...,X_{n+l}=x_{n} \right\} Pr{X1=x1,...,Xn=xn}=Pr{X1+l=x1,...,Xn+l=xn}
马尔可夫过程
P r { X n + 1 = x n + 1 ∣ X n = x n , . . . , X 1 = x 1 } = P r { X n + 1 = x n + 1 ∣ X n = x n } Pr\left\{X_{n+1}=x_{n+1}|X_{n}=x_{n},...,X_{1}=x_{1}\right\}\\=Pr\left\{X_{n+1}=x_{n+1}|X_{n}=x_{n}\right\} Pr{Xn+1=xn+1∣Xn=xn,...,X1=x1}=Pr{Xn+1=xn+1∣Xn=xn}
时不变马尔可夫过程
P
r
{
X
n
+
1
=
b
∣
X
n
=
a
}
=
P
r
{
X
2
=
b
∣
X
1
=
a
}
Pr\left\{X_{n+1}=b|X_{n}=a\right\}=Pr\left\{X_{2}=b|X_{1}=a\right\}
Pr{Xn+1=b∣Xn=a}=Pr{X2=b∣X1=a}
一般默认马尔可夫过程为时不变的
马尔可夫链表征
若{
X
i
X_{i}
Xi}为马尔可夫链,则
X
n
X_{n}
Xn为n时刻状态
时不变马尔可夫链可由初始状态和概率转移矩阵P=[
P
i
j
P_{ij}
Pij]表征
P
i
j
=
P
r
{
X
n
+
1
=
j
∣
X
n
=
i
}
P_{ij}=Pr\left\{X_{n+1}=j|X_{n}=i\right\}
Pij=Pr{Xn+1=j∣Xn=i}
有
p
(
x
n
+
1
)
=
∑
p
(
x
n
)
P
x
n
x
n
+
1
p(x_{n+1})=\sum p(x_{n})P_{x_{n}x_{n+1}}
p(xn+1)=∑p(xn)Pxnxn+1
平稳分布
n+1时刻与n时刻分布相同
若马尔可夫链初始状态平稳,则为平稳过程
有限状态不可约非周期的马尔可夫链趋于平稳分布
熵率
若极限存在
H
(
χ
)
=
lim
n
→
x
1
n
H
(
X
1
,
.
.
.
,
X
n
)
H(\chi)=\lim_{n\rightarrow x}\frac{1}{n}H(X_{1},...,X_{n})
H(χ)=n→xlimn1H(X1,...,Xn)
若为独立同分布
H
(
χ
)
=
H
(
X
1
)
H(\chi)=H(X_{1})
H(χ)=H(X1)
独立非同分布
H
(
χ
)
=
lim
n
→
x
1
n
∑
H
(
X
i
)
H(\chi)=\lim_{n\rightarrow x}\frac{1}{n}\sum H(X_{i})
H(χ)=n→xlimn1∑H(Xi)
平稳马尔可夫链熵率
H ( χ ) = H ( X 2 ∣ X 1 ) H(\chi )=H(X_{2}|X_{1}) H(χ)=H(X2∣X1)