Can MQL programmers be considered as programmers? - page 7
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
San Sanych, I'll let you in on a terrible secret: so is MQL. It's also an interpreter.
Where did you get the firewood?
San Sanych, I'll let you in on a terrible secret: so is MQL. Also an interpreter.
I'll join the queue and let you in on a no less terrible secret: ex4 and ex5 files are native code. ))
Useful information :
http://keldysh.ru/papers/2013/prep2013_19.pdf
M.A. Ananiev, N.A. Mitin
Comparison of Linear and Nonlinear Autoregressive Models of Conditional Heteroscedasticity Using RTS Index Return as an Example
ANNOTATION
In this paper we compare the forecasting capabilities of linear and non-linear conditional volatility models by the example of GARCH models for the RTS index return. Based on the daily closing prices of the RTS index for 10 years a set of parametric models is estimated and a set of volatility forecasts for different length horizons is built. The forecasting capabilities of the models are compared according to the selected criteria. Non-linear models have been developed to account for detected features of the time series, but the quality of the forecasts obtained with their help is sometimes questioned. The results of this study complement the results of other works: non-linear conditional volatility models show better results. A possible explanation for this success could be the fact that non-linear models give a better forecast over relatively short time horizons, while over longer time horizons they may give a larger error.
Of course, thank you.
But there's a huge literature on the application of Garch, and it's especially important in financial markets. There was an article somewhere that looked for garbage parameters on the example of all stocks in the S&P500 index, which is 500 stocks.
As I've read (not my experience, I can't repeat everything, it's too long), today the most advanced models are RealGARCH. The prefix Real refers to the variance already realised, i.e. the model uses two variances: on the larger TF and on the smaller TF for which there is a fact.
Everyone here is agitating for someone to start digging. I had such a companion, but he settled on an arima, which is part of the garch. And the amount of work is too much for me alone.
It is more convenient, not because the interpreter is secondary, but because R is a modelling environment, including (or primarily) statistical.
By the way, despite the fact that R is interpreted, the language itself is a scripting language and serves mainly for linking words in a sentence, i.e. functionality and various packages among themselves. And the language itself takes up a negligible part of the program execution time.
Thus, all complaints about the speed of R are completely unfounded. This is about using R directly in TC and senselessness of rewriting codes in MQL).
I completely agree with you: R is a very well thought out tool for research and development in statistics, and now machine modelling is also relegated to statistics. And it is very easy to use the results of research in industrial use.
I completely agree with you about the performance. In those algorithms that I use, I don't see any prospect of increasing the speed of the mcl enumeration.
And most importantly, I don't see such a need for rewriting at all - it's different tools for different areas combine perfectly and easily, everything works steadily.
The main thing is to learn GOTO and INPUT variables... on the ZX-Spectrum computer.
The rest is a waste of time.
The main thing is to learn GOTO and INPUT variables... on the ZX-Spectrum computer.
the rest is a godsend.
Hmm, in late 80's many people considered themselves programmers for their writing skills ))
LOAD ""Without that, though, you couldn't run any game (from a cassette recorder back then).
The problem is that pure GARCH(1,1) is a virtually unworkable model.
You have to take the appropriate package, the most interesting one is rugarch. You have to simulate the mean, ARCH proper, and there are plenty of these models, you can get good results with EGARCH, besides that you need to simulate the distribution. There are many publications highlighting the results of using this package in the financial markets including Forex. You can find here ready-made codes and examples, it's very instructive.
If you look at Rugarch and get a good result, it is available on Srp, the codes are open source.
But you're way off from Srp because it's not sure that you'll get a decent result with GARCH. In any case, it is much more convenient to conduct experiments in R rather than in µl, because R is an interpreter.
Useful information :
http://keldysh.ru/papers/2013/prep2013_19.pdf
M.A. Ananiev, N.A. Mitin
Comparison of Linear and Nonlinear Autoregressive Models of Conditional Heteroscedasticity Using RTS Index Return as an Example
ANNOTATION
In this paper we compare the forecasting capabilities of linear and non-linear conditional volatility models by the example of GARCH models for the RTS index return. Based on the daily closing prices of the RTS index for 10 years a set of parametric models is estimated and a set of volatility forecasts for different length horizons is built. The forecasting capabilities of the models are compared according to the selected criteria. Non-linear models have been developed to account for detected features of the time series, but the quality of the forecasts obtained with their help is sometimes questioned. The results of this study complement the results of other works: non-linear conditional volatility models show better results. A possible explanation for this success may be the fact that non-linear models give better forecasts at relatively short horizons, while at longer horizons they may give larger errors.
Hmm, in the late 80s many people considered themselves programmers for their writing skills )).
You really couldn't run any game without it (from a cassette recorder back then).
it's our own... and warm.
it's our FS)).
Hmm, in the late 80s many people considered themselves programmers for their writing skills )).
Without it, though, you couldn't run any game (from a cassette recorder back then).
I was 6 years old when I got my first Master as a gift.
But I don't remember writing Load ...