estimators by the method of moments and maximum likelihood?

1.1k Views Asked by At

Given this density function $f(x)=(\theta+1)x^{\theta}$ in $[0,1]$

Find the estimator by the method of moments for $\theta$ and by the maximum likelihood method....

I'm trying to use parametric method of moments but I guess very wrong. What is the right procedure?

1

There are 1 best solutions below

2
On

The method of moments estimator is found by taking the raw moments of the distribution and equating them with the sample moments, until a unique solution is found for the resulting system.

Here is an example using an exponential distribution: suppose $X \sim \operatorname{Exponential}(\lambda)$, where $$f_X(x) = \lambda e^{-\lambda x} \, \quad x > 0.$$ It is easy to calculate $$\operatorname{E}[X] = \int_{x=0}^\infty x f_X(x) \, dx = \int_{x=0}^\infty \lambda x e^{-\lambda x} \, dx = \frac{1}{\lambda}.$$ Now, if we are given an IID sample $$\boldsymbol X = (X_1, X_2, \ldots, X_n)$$ from this distribution, the sample mean is simply $$\bar X = \frac{1}{n} \sum_{i=1}^n X_i.$$ Therefore, the method of moments estimator for $\lambda$ arises by equating the first raw moment $\operatorname{E}[X]$ with the sample mean $\bar X$, and solving for $\lambda$: $$\tilde \lambda = \frac{n}{\sum_{i=1}^n X_i}.$$

In the case where you might want the joint method of moments estimators of a distribution with two unknown parameters, you might equate both the first raw moment $\operatorname{E}[X]$ and the second raw moment $\operatorname{E}[X^2]$ with the sample mean $\bar X$ (which is the first sample moment), and the mean sum of squares $$\overline{X^2} = \frac{1}{n} \sum_{i=1}^n X_i^2,$$ also known as the second sample moment, respectively. This is because if you have two unknown parameters, you would need to solve a system of two equations for the two unknowns, and using only the first moment will give only one equation or constraint.

In your case, you have only one parameter $\theta$, so all you need to do is compute the expectation of the distribution you are given, and equate it to the sample mean $\bar X$, and solve the resulting equation for $\theta$.