Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I wondering how to prove $X$, $Y$, $g(Y)$ is a Markov Chain in That Order?

$X$, $Y$, $Z$ is a Markov Chain in That Order (denoted $X\to Y\to Z$) if $$p(x,y,z) = p(x)\cdot p(y\mid x)\cdot p(z\mid y).$$ A property of a Markov Chain is if $X\to Y\to Z$, then $$ p(xz\mid y) = p(x\mid y)\cdot p(z\mid y).$$

Setting $Z = g(Y)$, I must show: $p(x,y,g(y)) = p(x)\cdot p(y\mid x)\cdot p(g(y)\mid y)$.

I believe I must simplify $p(g(y)\mid y)$ somehow, but I am only aware of conditional expectation identity where $E[g(Y)\mid Y] = g(Y)$ and $E[g(Y)\mid Y=y] = g(y)$.

Thanks for your assistance.

share|cite|improve this question
up vote 1 down vote accepted

We know that for any $X,Y,Z$:

$$p(X,Y,Z) = p(X) \cdot p(Y|X) \cdot p(Z|X,Y)$$

And it is a Markov Chain if:

$$p(Z|X,Y) = p(Z|Y)$$

Therefore, to show that $X,Y,g(Y)$ is a Markov Chain, we must show that:

$$p(g(Y)|X,Y) = p(g(Y)|Y)$$

This is trivial, because:

$$p(g(Y)=z|X=x,Y=y) = \begin{cases} 1 &\text{if } z = g(y), \\ 0 &\text{if } z \ne g(y) \end{cases}$$

That is, $g(Y)$ is uniquely determined by $Y$, which is to say, $X$ and $g(Y)$ and conditionally independent.

share|cite|improve this answer

We wish to prove if $Z = f(Y)$, then $X\to Y\to Z$.

We want to show: $p(x,y,z) = p(x)p(y\mid x)p(z\mid y)$, where $z=g(y)$. The following statement is always true:

$p(x,y,z) = p(x)p(y\mid x)p(z\mid yx)$. To reach the desired statement, we only need to show: $p(z \mid yx) = p(z \mid y)$.

$$ p(z \mid yx) = \begin{cases} 1, & z = g(y) \\ 0, & z \neq g(y)\end{cases}$$

Notice that $p(z \mid yx)$ does not depend on $x$. Then, $p(z \mid yx) = p(z \mid y)$ and we have proved $X \to Y \to Z$, when $Z = g(Y)$.

share|cite|improve this answer
Additional resource could be found here: – jrand Nov 8 '12 at 22:33

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.