Notation for the General Transformations of Random Variables?

31 Views Asked by At

For a continuous RV $X$ the general transformation to a RV $Y$ is given by:

$ P_y(y) = P(Y≤y) = P(f(X)≤y) = P(X \in \{x|f(x) ≤y\}) $

Could someone explain the meaning opf the final term please? This notation comes from the book Machine Learning: A Probabilistic perspective by Murphy on page 50.

When the functions are monotonic and invertible the no0tation becomes

$P_y(y) = P(Y≤y) = P(f(X)≤y) = P(X ≤ f^{-1}(y)) = P_x(f^{-1}(Y))$

which I understand.

1

There are 1 best solutions below

1
On BEST ANSWER

That line can be read as "the probability that $X$ is a value such that it's image (under $f$) is less than or equal to $y$" If our transformation isn't invertible, this is all the further we are able to simplify, unlike invertible transformations that can be further written as $P(X \le f^{-1} (y))$.

A good example would be to use the space $(-\infty, \infty)$ for $X$ and the transformation $Y = f(X) = X^2$. This transformation is not one-to-one and not invertible. In this case $f^{-1}(y)$ doesn't have a definite meaning.