Econometrics: let's discuss the CU balance sheet. - page 11
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Well, with mo=constant and variance=constant it is absolutely clear what the model should be and what the deterministic component should be. I.e. a linear trend.
No one is against a linear trend. The whole issue is the error of such approximation. I wrote about it above.
No it isn't. But if deviations from this wonderful trend introduce you to kolyan... We don't want to meet him here. That's what it's all about.
It's called: "sidetracking ;)))), we're not talking about kolyan now ;)))
What I'm saying is that the normality of the residuals is not a criterion for the quality of the TC.
Nobody is against a linear trend. The whole issue is the error of such an approximation. I wrote about it above.
I already wrote above that if Mo varies within a small range with finite variance (generally close to normal over a long period), then Mo calculated over the entire series will characterize the entire curve and the residuals should be distributed normally. And if the trend component (Mo) can rule as it pleases)), then why do we need such a system?
It's called: "sidetracking ;)))), we're not talking about kolyan now ;)))
What I'm saying is that the normality of the residuals is not a criterion for the quality of the TC.
And that's what I'm talking about. If the residual is normal, then everyone knows about the deviation from the average, and if not? That's why normality is a quality criterion, not profit value. If the distribution is normal, the profit value (mo) can be trusted, but if it is non-stationary, no testing can be trusted.
I already wrote above that if Mo varies within a small range with finite variance (generally close to normal in the long run), then Mo calculated over the entire series will characterize the entire curve and the residuals should be distributed normally. But if the trend component (Mo) can rule at will, then why do we need such a system?
Yes, probably. Everything I've written is suitable for kotir. And for balance..... I would like to go strictly upwards.
Yes, go on. Show the CU with an even distribution of the residuals.
It goes like this.
Detrending a straight line. The slope angle is determined by the variance so as not to catch a margin call. The variance can change, but stationary, or better still, so that predictions can be made.
That's not the point... You can make pictures on the subject, if you're interested... The point is that the normality of the residuals is not a defining feature of TC quality, but can only be regarded as a secondary indicator. That is, of course, assuming that the goal of the TS is to grow profits and not to flatten the plaques on the balance line.
Here you have a series of equities. You don't know anything about the trend component about the balances, etc. What are your requirements for it?
The point is that the normality of the residue is not a determining sign of the quality of TC, but can only be regarded as a secondary indicator. That is, of course, assuming that the goal of the TS is to grow profits and not to align the plaques on the balance line.
The point is that the balance is primary.
If the balance is normal or stationary (as it seems to me), then we can talk about the test results: we can discard a loss-making TS and keep a profitable one. But if the balance is not stationary, we cannot say anything about the TS and it does not matter if it is profitable or not during testing - it does not exist at all.
Well, I replied to the automaton about it above.
So, you write:
"A model without normality will be correct and adequate with a certain accuracy if the series is stationary. "
If the series is stationary, then by subtracting the trend (mo), the residual will be normal. I.e. the residual analysis is the assessment of robustness or stationarity of the distribution (which is actually the same thing).
P.S. the first differences are stationary and the equity series itself has a unit root
Where does it come from? From what does it follow? Few bell-shaped distributions other than normal?
What does this have to do with first differences? Is the equity series has a unit root, or is that a Kama Sutra lesson?