MHB Calculate the density - Unbiased estimator for θ

  • Thread starter Thread starter mathmari
  • Start date Start date
  • Tags Tags
    Density
mathmari
Gold Member
MHB
Messages
4,984
Reaction score
7
Hey! :giggle:

For $n \in \mathbb{N}$ we consider the statistical product model $(X ,(P_{\theta})_{\theta\in\Theta})$ with $X = (0,\infty)^n$, $\Theta = (0,\infty)$ and densities $f_{\theta}(x_i) = \frac{1}{\theta} \textbf{1}_{(0,\theta)}(x_i)$ for all $x_i \in (0,\infty)$, $\theta \in \Theta$. It will be the three estimators $T$, $\tilde{T}$ , $\hat{T}$ : $X \rightarrow \mathbb{R}$, $$T(x)=\frac{2}{n}\sum_{i=1}^nx_i, \ \ \ \tilde{T}(x)=c\cdot \max (x_1,\ldots ,x_n), \ \ \ \hat{T}(x)=2x_1$$ proposed for $\theta$, where $c \in (0,\infty)$ is a constant.

(a) Show that $f : \mathbb{R} \rightarrow [0,\infty)$, $f(y) = \frac{n}{\theta^n} y^{n-1}\textbf{1}_{(0,\theta)}(y)$, is the density of $\max(X_1, \ldots , X_n)$ by first calculating the distribution function $F(y) = P_{\theta}[\max(X_1,\ldots , X_n) \leq y]$ and then deriving it.

(b) For which $c \in (0,\infty)$ is $\tilde{T}$ unbiased for $\theta$ ?

(c) Calculate for $T$, $\tilde{T}$ and $\hat{T}$ each the mean square deviation at $\theta$.
At (a) we have :

We have that $\max(X_1,\ldots , X_n) \leq y$ if $X_i\leq y$ for all $1\leq i\leq n$, right?
Therefore we get $F(y) = P_{\theta}[\max(X_1,\ldots , X_n) \leq y]=\left (P_{\theta}[X_1\leq y]\right )^n$.
Do we know that probability? How could we continue? :unsure: At (b) do we check if $E(T)=\theta$ , $E(\tilde{T})=\theta$ and $E(\hat{T})=\theta$ ? Or what are we supposed to check ? :unsure:
 
Physics news on Phys.org
Hi mathmari,

Good work so far. Here are a few ideas to keep things moving.

(a) You're correct that $\max(X_1,\ldots , X_n) \leq y$ iff $X_i\leq y$ for all $1\leq i\leq n$. It is also true in this case that $F(y) = \left (P_{\theta}[X_1\leq y]\right )^n$, but this does not follow just because $\max(X_1,\ldots , X_n) \leq y$ iff $X_i\leq y$ for all $1\leq i\leq n$. The reason this is true is because each of the variables $X_{i}$ is distributed identically/the same way (in this case uniformly on the interval $(0,\theta)$) according to the probability distributions/densities $f_{\theta}(x_{i})$. If this were not the case, you could not conclude $F(y)=\left (P_{\theta}[X_1\leq y]\right )^n$.

In general (i.e., this applies broadly and not just to this problem), to obtain a probability from a density, we integrate the density over an appropriate interval. For example, if we wanted to know the probability that $0.25\leq X_{i}\leq 0.75$, we would calculate $$\int_{0.25}^{0.75}f_{\theta}(x_{i})dx_{i}.$$
Try using this information to determine what $P_{\theta}[X_1\leq y]$ is. The cumulative probability is then $F(y) = \left (P_{\theta}[X_1\leq y]\right )^n.$ By definition of probability densities (see Wikipedia - Probability Density), the probability density of $\max(X_1,\ldots , X_n) $ is the derivative of $F(y)$ with respect to $y$.

(b) You need to determine the value of $c$ so that $\theta = E(\tilde{T})$. In general, if a random variable $X$ has a probability density function $f(x)$, then the expected value of $X$ is $$E[X] = \int_{-\infty}^{\infty}x\cdot f(x)dx$$
You will want to use your answer from part (a) to calculate $E(\tilde{T})$.

Feel free to let me know if you have any other questions.
 
Last edited:
Hi all, I've been a roulette player for more than 10 years (although I took time off here and there) and it's only now that I'm trying to understand the physics of the game. Basically my strategy in roulette is to divide the wheel roughly into two halves (let's call them A and B). My theory is that in roulette there will invariably be variance. In other words, if A comes up 5 times in a row, B will be due to come up soon. However I have been proven wrong many times, and I have seen some...
Thread 'Detail of Diagonalization Lemma'
The following is more or less taken from page 6 of C. Smorynski's "Self-Reference and Modal Logic". (Springer, 1985) (I couldn't get raised brackets to indicate codification (Gödel numbering), so I use a box. The overline is assigning a name. The detail I would like clarification on is in the second step in the last line, where we have an m-overlined, and we substitute the expression for m. Are we saying that the name of a coded term is the same as the coded term? Thanks in advance.
Back
Top