Reading the following paragraph I was wondering how this entity that the authors [1] call a sub-vector should be named. In Matlab a sub-vector has to be contiguous.

Let $\mathbf{X}=(X_1,\ldots,X_n)$ be a vector of random variables and $\mathbf{x}=(x_1,\ldots,x_n)$ a possible value setting (configuration) for these variables. $x_i$ denotes a possible value of $X_i$, the $i$th component of $\mathbf{X}$, and $\mathbf{y}$ denotes a possible value setting for the sub-vector $\mathbf{Y}=(X_{J_1},\ldots,X_{J_k})$, $J=\{J_1,\ldots,J_k\}\subseteq\{1,\ldots,n\}$.

Is sub-vector indeed the proper mathematical term for this entity?

[1] A review of probabilistic graphical models in evolutionary computation [Larrañaga et al.], 2012

  • $\begingroup$ I haven't heard it before, but this 'sub-vector' seems appropriate term for that, I would accept and use it without hesitation.. $\endgroup$ – Berci Mar 12 '13 at 13:44
  • $\begingroup$ You can call $\mathbf{X}$ a sequence, and then $\mathbf{Y}$ is a subsequence, which is standard. But when it can't cause any confusion, sub-vector sounds like a reasonable name too. $\endgroup$ – Alfonso Fernandez Mar 12 '13 at 13:52
  • $\begingroup$ From the definition it doesn't seem that the order needs to be preserved, so sub-vector seems to be - indeed - more general than subsequence. I am not a mathematician, but has the prefix "sub" anything to do with a subspace that can be defined? Or is it free to use? $\endgroup$ – Anne van Rossum Mar 14 '13 at 9:22
  • $\begingroup$ @AnnevanRossum 'sub-' is used in a lot of different ways (subgroup, subgraph, subword, subspace), it just generally means a 'smaller' version of an object that has the same kind of structure. $\endgroup$ – Jair Taylor Aug 2 '18 at 22:00

"Subvector" doesn't seem to have a standard definition. As you've noted, it's different in Matlab.

According to some authors (eg. Dominic Welsh in his book Matroid Theory), a subvector is a vector that's in the same general direction as another vector, but shorter. To be more precise, let $u = (u_1, u_2, \dots, u_n)$ and let $v = (v_1, v_2, \ldots, v_n)$. If each $v_i - u_i$ is nonnegative, then $u$ is a subvector of $v$.

The conclusion is that if you're going to use the term "subvector", it's best to define it first.


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.