2
$\begingroup$

Suppose the joint survival function of the latent failure times for two competing risks, $X$ and $Y$, is $S(x,y)=(1-x)(1-y)(1+0.5xy)$, $0<x<1$, $0<y<1$. Find the cumulative incidence function of $X$?

I first solved the marginal cumulative distribution function of $X$: $(1-x)$. Then I tried to find the joint density function: $1.5-x-y+2xy$, but I am unable to determine how to properly integrate this to find the cumulative incidence function of $X$.

$\endgroup$
0
$\begingroup$

For $0 \leq x \leq 1$ the cumulative incidence of $X$ is defined as $$ \mathbb P \left (X \leq x, X \leq Y \right) $$

To compute this probability we need to integrate the joint density of $(X,Y)$,

$$ f(x,y) = \frac{3}{2} -x-y+2xy $$

over the set $\mathcal{A} \equiv \{(u,v) \in [0,1]^2 \mid u \leq x \wedge u \leq v \} $

That is,

\begin{align*} \mathbb P \left (X \leq x , X \leq Y \right) &= \int_\mathcal{A} f(u,v)\text{d}u\text{d}v \\ &= \int_0^x \left( \int_u^1 f(u,v)\text{d}v \right )\text{d}u \\ &= \int_0^x \left( \int_u^1 \left(\frac{3}{2} -u-v+2uv \right)\text{d}v \right )\text{d}u \\ &= \int_0^x \left( \frac{3}{2}(1-u) -u+u^2 - \left[ \frac{v^2}{2} \right]_u^1 +2u \left[ \frac{v^2}{2} \right]_u^1 \right) \text{d}u\\ &=\int_0^x \left( \frac{3}{2} - \frac{3}{2} u -u+u^2 - \left( \frac{1}{2}-\frac{u^2}{2} \right) +2u \left( \frac{1}{2} - \frac{u^2}{2} \right) \right) \text{d}u\\ &=\int_0^x \left( 1- \frac{3u}{2} + \frac{3u^2}{2} -u^3\right) \text{d}u\\ &= \left[ u - \frac{3u^2}{4} + \frac{u^3}{3} - \frac{u^4}{4} \right ]_0^x \\ &= x - \frac{3x^2}{4} + \frac{x^3}{3} - \frac{x^4}{4} \\ &= \frac{12x -9x^2 + 6x^3 - 3x^4}{12}. \end{align*}

A quick check, take $x=1$ we have $$ \frac{12x -9x^2 + 6x^3 -3x^4}{12} = \frac{1}{2} $$ and $$ \mathbb P \left (X \leq 1, X \leq Y \right) \stackrel{(1)}{=} \mathbb P(X\leq Y) \stackrel{(2)}{=}\frac{1}{2} $$

Here $(1)$ comes from the fact that since $X \leq 1$ with probability $1$, the event $\{X \leq 1 \cap X \leq Y\}$ has the same probability as the event $\{X \leq Y\}$ and since $X$ and $Y$ are somehow "symmetric" we have $\mathbb P(X\leq Y) = \mathbb P(Y\leq X) = \frac{1}{2}$ hence equality $(2)$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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