Suppose I have $N$ continuous independent random variables (random vectors) defined on $\mathbb{R}^N$. Can I comment on the probability of a particular realization of these N vectors being LINEARLY independent? In particular, can I comment that $P(\sum_i \alpha_i X_i=0) = 0$ for any set of $\alpha_i\in \mathbb{R}$ (expect all zeros) and where $X_i$'s are the $N$ random vectors.

  • $\begingroup$ We need to know something about the distribution of this random vectors. With a Dirac delta distribution, it is also possible that such probability is one. $\endgroup$ – Jack D'Aurizio Sep 15 '14 at 9:51
  • $\begingroup$ Since they are continuous, Dirac's are excluded. $\endgroup$ – Ajay Sep 15 '14 at 11:31

For a continuous distribution of $\mathbb{R}^n$, the measure of a $d$-dimensional vector space $V\subset\mathbb{R}^n$ with $d<n$ is zero, hence $$\mathbb{P}[X_n\in\operatorname{Span}(X_1,\ldots,X_{n-1})]=0,$$ so you are right.

  • $\begingroup$ Thanks. I believe you meant $P[X_n \in \text{Span}(X_1,\cdots, X_{n-1})]=0$ $\endgroup$ – Ajay Sep 15 '14 at 14:04

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.