Econometrics: one step ahead forecast - page 37
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
1. (18) works and makes me a profit, show a similar one in this area.
Can't you upload an rar archive to ifolder? Interesting. Downloaded it, renamed the extension, it all opened, thanks.
It's very difficult to evaluate it in an hour, the book is global. It's a very serious guide to econometrics. That's what I'm going to study it on. The presentation is quite simple, with plenty of illustrations and applications for finnings.
P.S.
In what follows, we use S-Plus in empirical illustrations. Other software packages (e.g., Eviews, SCA, R, and RATS) can also be used.
That's not so good, but I guess you can get used to it.
And here is the book itself. Unfortunately I only found the translation. I translated it myself. Maybe it will help someone. ECONOMETRY
It is very difficult to evaluate it in an hour, the book is global. A very serious guide to econometrics. That's how I will study it. The presentation is simple enough, there are many illustrations and applications for finn cials.
I would like to share my experience.
A year ago I saw EViews. Before that I knew 90% of econometrics, but it just did not stick in my head. After a year of dealing with EViews something started to come together, into a certain sequence of actions, which resulted in obtaining quite decent trading systems. But EViews has taught that you can't believe anything - you have to have proof that you can use it and it's not just and not so much forward tests.
For example, you correctly noted above that R-squared cannot be completely trusted. But those who use a thousand indicators have ever heard of R-square?
The R-square is important, but apart from that there are many other things.
If we look at EViews - it is a lot of tools that you do not know how and when to apply, and when applied - you do not know how to assess the result.
I have prepared this topic with two articles and a code on EViews and MQL4 for some reason.
At present I have a certain opinion about a number of econometric tools and EViews which are sufficient to build a decent TS. But I want to clarify my notions and expand them, if possible. The space of states is very interesting. Automat promised to give a model after the weekend.
I am not going to teach anyone econometrics, much less defend it.
I outlined my plan in the beginning of the topic: I propose a model, demonstrate its results and propose:
a) discuss these results
b) Modernize the model.
I am currently using model #2. I know that it is not workable. Let's improve it. There is a specific idea in building the model: isolate the trend and account for the noise.
http://ifolder.ru/27037398
It's rocking all right.
More than decent book for a first read
Basic book
Hamilton. Time Series Analysis
Available online. 23 MB
Nosko Econometrics - Introduction to Regression Series Analysis
Nosko Econometrics for Beginners
Nosko Econometrics for Beginners Supplementary Chapter
Nosko Econometrics for Beginners Supplementary Ch.
All books are publicly available.
Closing the forecast for Friday on Close. Here is the result:
Some conclusions:
1. The DX prediction is much better than the lagged EURUSD itself
2. The results of the forecast are qualitative (matched - not matched) and do not take into account the MM and the spread. For example, on the last day, Friday, calculated price = 1.3514 and High = 1.3613. When using the DX forecast the potential profit was 100 pips higher. On the other hand, Low=1.3447, and using an unsuccessful forecast on EURUSD using the sliding trawl for SL, the loss would have been minimal.
3. The table presented cannot be the basis for using the model due to the small sample size. The need to use a tester is obvious to all. Such a possibility is available. The corresponding code is laid out in the attachment to my article. But I will not do it, as in my opinion the model is not ready and needs to be finalised before final testing.
My plan is as follows:
1. I finish making predictions.
2. I suggest that everyone who is interested:
a) discuss these results
b) modernise this model.
c) propose their models
3. I am willing to implement the results of discussions and upgrades in code and post the results.
Let me remind you of the type of models:
a) For EURUSD on lags: EURUSD = hp(-1 to -4) + hp_d(-1 to -2)
b) For DX:
DXM = 1/DX - we use the inverse of the quotient
EURUSD = DXM_HP(-1 TO -4) + DXM_HP_D(-1 TO -2)
In these formulas HP is the Hedrick-Prescott indicator, and HP_D is the residual = kotir - indicator. The bars in brackets are the bars before the current one, (-1 to -4) means the last 4 bars.
The actual equation after evaluating the coefficients at the variables is as follows:
EURUSD = -1552.7613734*DXM_HP(-1) + 4731.89082764*DXM_HP(-2) - 4360.68995095*DXM_HP(-3) + 1287.82064375*DXM_HP(-4) - 98.9244837504*DXM_HP_D(-1) - 131.011472103*DXM_HP_D(-2)
Those who wish can implement this formula in the form of an indicator, about which it will be known that corresponds to a quote of 97%. A very high quality mash-up!
Closing the forecast for Friday on Close. Here's the result:
Some conclusions:
1. The forecast by DX is much better than the lag values of EURUSD itself
My plan is as follows:a) For EURUSD on lags: EURUSD = hp(-1 to -4) + hp_d(-1 to -2)
b) For DX:
DXM = 1/DX - we use the inverse of the quotient
EURUSD = DXM_HP(-1 TO -4) + DXM_HP_D(-1 TO -2)
Don't rush to close the thread as a day or two forecast is very little for conclusions
The DXM model is more realistic according to your forecast.
Still some way to go to the code and run it through history as the result is interesting?
Question: what is DXM_HP(-1 TO -4) - DXM_HP function for last 4 bars or what? Also what TF?
Or have you already decided that you have told too much and achieved positive results?
I'm interested in this thread so far as I still don't understand - is the forecasting based on statistics? If so, I'm off)))
new-rena:
Don't rush to close the thread, as a forecast for a day or two is very little for conclusions
The branch is not closed. But I opened it with a specific purpose: to collectively deal with the technology of TC development using generally accepted science.
Still, is there any way to go to the code and run through the history, as the result is interesting?
A set of statistics on a particular model does not interest me - I already know that it does not work. And this knowledge is based not on tester results, not on forward tests, but on known to me defects in the internal structure of the model. Nothing like TA gives you anything. Only confidence in the correctness of the model design, confirmed by tests of course, can give a basis for profitable trading. Without that any testing is complacency.
Question: what is DXM_HP(-1 TO -4) - is it the DXM_HP function for the last 4 bars or what?
The formula is deciphered below.
Also what is the TF?
D1 - you can see it from the results table.
I'm interested in this thread so far as I still don't understand - is the forecasting based on statistics? If so, I am gone)))
Yes, other methods, for example all TA. which do not calculate at least prediction error, are not recognized by me and I refer them to alihi, astrology and other "sciences".
faa1947:
2. I invite everyone to:
(a) Discuss these results