Transformation of Random variables

Function of One Random Variable

Given that Y ( ) = g [ X ( )] and the probability distribution of X find the probability distribution of X , we would like to find the probability distribution of Y . By definition
FY ( y ) = P{Y ( ) y} = P{g ( X ( )) y}.

Since the statistics of X ( ) is known, the last term, namely, P{g ( x ) y} could be determined in terms of y . When FY ( y ) is known, then the density function may be found. That is,
fY ( y) = dFy ( y ) dy

Fundamental Transformation Theorem

Given f X (x ) and Y = g ( X ) , then as the probability density function Y is given as fY ( y) =
i =1 n

f X ( xi ( y )) , g ( xi ( y ))

where xi = g 1 ( y ) are n real roots for a given y . If for some value of y there is no real root, then
fY ( y) = 0 .

Justification: By definition,
f Y ( y )dy = P{y < Y y + dy}

Suppose for a given y there are n roots, i.e.

y = g ( xi ) , xi ~ root, i = 1,2,..., n

Thus f Y ( y )dy = P{x1 < X < x1 + dx1 ... x n < X x n + dx n }. or

f Y ( y )dy = f X (xi ) dxi .
i =1 n

Therefore,
fY ( y ) =
i =1 n

f X ( xi ) n f X ( xi ) = . dy i =1 g ( x i ) dxi