Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. 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

Random variable $X$ has the probability density function \begin{equation*} f\left( x\right) =\left\{ \begin{array}{ccc} n\left( \frac{x}{\theta }\right) ^{n-1} & , & 0<x\leqslant \theta \\ n\left( \frac{1-x}{1-\theta }\right) ^{n-1} & , & \theta \leqslant x<1% \end{array}% \right. \end{equation*} show that if $k\in \mathbb{N}$ \begin{equation*} \mathrm{E}\left( X^{k}\right) =\frac{n\theta ^{k+1}}{n+k}+\sum% \limits_{i=0}^{k}\left( -1\right) ^{i}\binom{k}{k-i}\frac{n}{n+i}\left( 1-\theta \right) ^{i+1} \end{equation*}

It is easy to find the first term of $\mathrm{E}\left( X^{k}\right) $ but i couldn't find the second one. i think i have to use beta distribution properties. i tried to simulate the integral \begin{equation*} \int\nolimits_{\theta }^{1}nx^{k}\left( \frac{1-x}{1-\theta }\right) ^{n-1}% \mathrm{d}x \end{equation*} to beta pdf using $u=\frac{x-1}{\theta -1}$ transformation but i couldn't get a reasonable result. After this transformation is applied i have to find \begin{equation*} \int\nolimits_{0}^{1}u^{n-1}\left[ 1+u\left( \theta -1\right) \right] ^{k}% \mathrm{d}x \end{equation*} but i couldn't. Also I tried to use the equality \begin{equation*} x^{k}=1+\left( x-1\right) \sum\limits_{n=0}^{k-1}x^{n} \end{equation*} but i couldn't get the result.

share|improve this question
up vote 9 down vote accepted

$$\begin{align*} {\rm E}[X^k] &= \int_{x=0}^\theta x^k n \biggl(\frac{x}{\theta}\biggr)^{\!n-1} \, dx + \int_{x=\theta}^1 x^k n \biggl(\frac{1-x}{1-\theta}\biggr)^{\!n-1} \, dx \\ &= \frac{n}{\theta^{n-1}} \int_{x=0}^\theta x^{n+k-1} \, dx + \frac{n}{(1-\theta)^{n-1}} \int_{x=0}^{1-\theta} (1-x)^k x^{n-1} \, dx \\ &= \frac{n}{\theta^{n-1}} \cdot \frac{\theta^{n+k}}{n+k} + \frac{n}{(1-\theta)^{n-1}} \int_{x=0}^{1-\theta} \sum_{i=0}^{k} \binom{k}{i} (-x)^i x^{n-1} \, dx \\ &= \frac{n \theta^{k+1}}{n+k} + \frac{n}{(1-\theta)^{n-1}} \sum_{i=0}^{k} (-1)^i \binom{k}{k-i} \int_{x=0}^{1-\theta} x^{n+i-1} \, dx \\ &= \frac{n\theta^{k+1}}{n+k} + \frac{n}{(1-\theta)^{n-1}} \sum_{i=0}^k (-1)^i \binom{k}{k-i} \frac{(1-\theta)^{n+i}}{n+i} \\ &= \frac{n\theta^{k+1}}{n+k} + \sum_{i=0}^k (-1)^i \binom{k}{k-i} \frac{n}{n+i} (1-\theta)^{i+1}. \end{align*}$$

share|improve this answer
    
(+1) The second integral is (by inspection) an incomplete Beta function: knowing this makes further analysis possible, makes many other formulas immediately accessible, and also leads to more efficient computation (when $k$ is medium to large). – whuber Apr 16 '14 at 15:08

Your Answer

 
discard

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.