Let $(x_1, x_2,\dotsc, x_n)$ be a sequence of vectors: ($\forall i=1,\dots,n$) \begin{pmatrix} x_i^1 \\ x_i^2\\ \vdots\\ x_i^m \end{pmatrix}

In statistics, one often has to compute the sample mean vector: $$X = \frac{1}{n}\sum_{i=1}^n x_i$$

and the sample covariance (or variance-covariance) matrix: $$\mathbb{V} = \frac{1}{n-1}\sum_{i=1}^n (x_i - X)(x_i - X)^T$$ where $x^T$ is the transpose of $x$.

My questions are:

  1. what canonical form of information would I suggest to represent the sequence $(x_1, x_2, \dotsc, x_n)$ in order to compute the sample mean vector and the sample covariance matrix?

  2. How can I verify that all the desirable properties of canonical information are satisfied: Existence and Uniqueness, Completeness, Elementary, Empty, Combination, Update, and Compactness and Efficiency.

  3. What are the minimum number of observations $n$ for which $X$ and $\mathbb{V}$ are defined?

Thanks in advance.

  • $\begingroup$ Please show your work. $\endgroup$ – PackSciences Feb 19 at 16:45
  • $\begingroup$ @PackSciences Unfortunately, I don't even have an idea of how to start this task. Could you help me, please? $\endgroup$ – Fouzi TAKELAIT Feb 22 at 0:51

Your Answer

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

Browse other questions tagged or ask your own question.