When I write something like
x is a stochastic function of y: x ~ N(2y, 3) (1),
how do I refer back to that... non-equation? Do I still put "see Eq. (1)" although (1) is not an equation?
Mathematically you are true... it's not an equation. Technically I would say, that everything, that is display-style math and has a number attached for identification is called an equation in such contexts.
It seems that the right thing to do will depend on what's conventional in your field. In pure mathematics, it's standard to refer to all displayed equations, inequalities, etc. using just numbers in parentheses (for example, "using (2.1)", without specifying there whether (2.1) is an equation). You could add a descriptive noun if you'd like to emphasize it (e.g., "using equation (2.1)"), but you don't need to. If you do add a noun, it could be considered strange to refer to anything but an actual equation as an equation. This style of referencing displays creates no ambiguity, since citations use square brackets and all other numerical references have an attached noun or symbol to indicate whether they refer to a theorem, section, etc.:  is a citation, (1) is an equation or other display, Lemma 1 is a lemma, Section 1 is a section, etc.
The system described in the previous paragraph presumably doesn't apply to the author of the question, since abbreviations like "Eq. (1)" or "Ineq. (1)" are not standard in pure mathematics, which suggests he is in another field. However, it's worth keeping in mind that conventions vary between fields, so there won't be an absolute answer to this question. To know for sure what would look reasonable, it's important to know the context (i.e., the audience for the paper and where it might be published).
If you absolutely don't want to use Eq. (1) for mathematical reasons (~ is not equal), I would suggest
See Formula (1)
although this is not conventional.
Someone suggested "See (1)". I personally would not use it because it has some ambiguity to me. Does it mean "See Sec. (1)" or something else?
As mentioned in some other answers, and with some addendum. Five examples of the same text in different styles (please, excuse my English):
We may now apply (15) to (13). From (12) we then see that (14) is satisfied.
We may now apply Eq. (15) to Eq. (13). From Eq. (12) we then see that Eq. (14) is satisfied.
We may now apply Eq. 15 to Eq. 13. From Eq. 12 we then see that Eq. 14 is satisfied.
We may now apply Equation (15) to Equation (13). From Equation (12) we then see that Equation (14) is satisfied.
We may now apply Equation 15 to Equation 13. From Equation 12 we then see that Equation 14 is satisfied.
(You can substitute "Eq." by "Ineq." etc., whatever you want.)
Such chains of links to equations are much common than for figures or tables. Now tell me in Examples 2 and 3 where the sentences stard and end. Yes, after a while, one sees that the sentence ends after "(13)", but it takes a lot of time to realize that. In Examples 4 and 5, the text gets unnecessarily long.
In my opinion, equations should be refered solely by their number in parentheses, references of course solely in brackets. For figures, enumerated lists, examples, sections, theorems etc., one should spell out the name (abbreviated or not, that's a personal taste) and add the number without any parentheses, even if it originally had some.
So a numbered list: "(1) apple; (2) banana" is still refered as: "In Item 2 we see that banana is a banana." If you refer items a lot, it's worth giving them a style that doesn't clash with the one for equations, like (a), (b), (c), ... or (i), (ii), (iii); then you can refer the items without the word "Item".
Some solutions of (3) were obtained by Doe in ; we list them in Table 5 and they are plotted in Figure 1.