1
$\begingroup$

Let's say I have a process $X_t$ with unknown variance process $V_t$.
Then, I write $\mathrm{EMA}[X_t]$ to be the 5 sec exponential moving average of $X_t$.

Consider the transformation $$\sum (X_t-\mathrm{EMA}[X_t]).$$

What can we say about the transformation in terms of the original $X_t$ and $V_t$? (Doesn't have to be proof, but I do like proofs)

As far as I know, a process minus it's exponential moving average is equivalent to fractional differencing. Summing the differential produces a process that is highly correlated to X, and has a more normalized (smoothed) variance process.

If you could point me in the right direction to study something like this I would be very appreciative.

Improve this question
$\endgroup$
4
  • $\begingroup$ Hi: I can't help but where did you get the info that a process minus it's EWMA is equivalent to fractional differencing. I'm not saying that's wrong but rather I'm just wondering where you got that from ? $\endgroup$ Commented Feb 22, 2020 at 12:01
  • $\begingroup$ Hi Mark. I am just trying to understand what theories apply to a process transformation like this. We could assume that the mean and variance are constant for X. A colleague informed me that fractional differencing is relevant for a problem like this, though I am unfamiliar with that set of techniques. $\endgroup$ Commented Feb 22, 2020 at 20:38
  • $\begingroup$ Hi: are you trying to do someting with financial returns ? if so, email me off list and I can send you something tha could be relevant ? If not, then I don't think I can help because, taking a process and subtracting the ewma from it, doesn't ring a bell to me except for this one specific paper I'm thinking of. Actually, I don't have time this second, but tonight I'll try to find it and send the link so that anyone can read it if they're interested. $\endgroup$ Commented Feb 23, 2020 at 21:49
  • $\begingroup$ Hi: Like I said, if you're analyzing returns ( keep in mind that an EWMA is always equivalent to a moving average of window length $n$ for $\lambda = \frac{2}{n+1}$ ), then the paper at this link could be useful. papers.ssrn.com/sol3/papers.cfm?abstract_id=2604942. Other than this possible relevance, I've never heard of the ewma being subtracted from a time series process which is why I asked where you got that from. It could be true for sure. I just don't understand it and have never come across it. Thanks. $\endgroup$ Commented Feb 23, 2020 at 23:54

0

Reset to default

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.