Conditional density function in a Pareto distribution with a condition on the parameter

728 Views Asked by At

Let $X$ be a random variable having density function $$f(x;\vartheta,\lambda)=\lambda\,\vartheta^\lambda\,x^{-(\lambda +1)}$$ $$\vartheta>0\qquad\lambda>0\qquad x\geq\vartheta$$ Ginen $\psi>\vartheta$, find the density funtion of $X$ conditioned to the event $\lbrace{ X>\psi\rbrace}$.

Intuitively, I'd say $$f(x;\vartheta,\lambda | \lbrace{ X>\psi\rbrace}) =\lambda\,\psi^\lambda\,x^{-(\lambda +1)}$$ $$\psi>0\qquad\lambda>0\qquad x\geq\psi$$

This can be found forcing $F(\psi)=1-\left(\frac{\vartheta}{\psi} \right)^\lambda = 0$, where $F(x)$ is the cumulative distribution function of the variable $X$.

Is this way of reasoning right or am I missing something? Thanks in advance for your help!

1

There are 1 best solutions below

1
On BEST ANSWER

Your intuition is correct, but you do need to formalize it. To do this efficiently, it is easiest to work with the survival function $$S_X(x \mid \vartheta, \lambda) = \Pr[X > x] = \int_{t=x}^\infty \lambda \vartheta^\lambda t^{-(\lambda+1)} \, dt = (\vartheta/x)^\lambda, \quad x \ge \vartheta.$$ So for $x \ge \psi \ge \vartheta$, $$\Pr[X > x \mid X > \psi] = \frac{S_X(x \mid \vartheta, \lambda)}{S_X(\psi \mid \vartheta, \lambda)} = \frac{(\vartheta/x)^\lambda}{(\vartheta/\psi)^\lambda} = (\psi/x)^\lambda = S_X(x \mid \psi, \lambda).$$ This completes the proof.